The European Business Review - January-February 2024

The
Revisiting HR Challenges
at the Dawn of 2024
......................................................
Why the Ethical Use of
European
AI Matters for Your Career
......................................................
The Passive-Aggressive Individual
......................................................
Edge Computing in Europe: A Key
Business
Driver of Business Innovation
Review
January - February 2024
europeanbusinessreview.com
A FORWARD-THINKING CONCEPT BY
HAMILTON MANN
FOUNDER AND HOST OF
The Hamilton Mann’s Digital
For Good Conversation
INTRODUCING THE CONCEPT

OF ARTIFICIAL INTEGRITY
THE PATH FOR THE FUTURE OF AI
USA $22 EU €17.5

empowering communication globally CAN $22 UK £15
Secure password
management
for business
with end-to-end encryption
The European
Business Review
empowering communication globally
JANUARY – FEBRUARY 2024
ARTIFICIAL INTELLIGENCE AI AND ETHICS

4 Introducing the Concept of Artificial 50 Why the Ethical Use of AI Matters for Your Career
Integrity: The Path for the Future of AI Jack McGuire, David De Cremer,
Hamilton Mann Leander De Schutter, and Yorck Hesselbarth
18 The Path towards Trustworthy AI is BUSINESS INNOVATION

no Tech but a Human Intelligence Test 54 Edge Computing in Europe:
Hamilton Mann A Key Driver of Business Innovation
Ram Ramalingam, Teresa Tung,
24 Enhancing Operating Models’ Artificial Nitu Kaushal, and Shalabh Kumar Singh
Intelligence Quotient (AIQ)
Hamilton Mann INNOVATION
60 Innovation, the art of abolishing distance:
32 Getting Closer to Machines with Building the ATTRACT DeepTech Ecosystem
Mindful Steps Hervé Legenvre
Robb Wilson
STRATEGY AND MANAGEMENT
38 The Key Success Factors of a 70 Why We Need a Resolution Revolution
Powerful AI Factory David Liddle
Jacques Bughin
PSYCHOLOGY
42 2024: What is the Near Future of 76 The Passive-Aggressive Individual
Generative AI? Adrian Furnham
Jacques Bughin
HR TRENDS
94 Revisiting HR Challenges at the Dawn of 2024
Dr. Pedro Cesar Martínez Moran and
Dr. Simon L. Dolan
Production Accounts: Lynn Moses. Head of Design and Production: Kimberly Barrera. Production & Design: Jenya Shliepova Editors: Elenora Elroy, David Lean. International Media
Editors: Ariane Cornejo, Maria Carmela Matibag. Editorial and Marketing: Pamela Martinez, Mary Celu Aratas. Digital Content: Angela Lebrino. Print Strategy: Stefan Newhart. Group
Managing Editor: Jane Liu. Editor in Chief: The European Business Review Publishing Oscar Daniel. READERS PLEASE NOTE: The views expressed in articles are the authors’ and not
necessarily those of The European Business Review. Authors may have consulting or other business relationships with the companies they discuss. The European Business Review:
3 - 7 Sunnyhill Road, London SW16 2UG, Tel +44 (0)20 3598 5088, Fax +44 (0)20 7000 1252, info@europeanbusinessreview.com, www.europeanbusinessreview.com No part of this
publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval
system, without written permission. Copyright © 2024 EBR Media Ltd.All rights reserved. ISSN 1754-5501
empowering communication globally

ARTIFICIAL INTELLIGENCE
The concept of “artificial integrity”

proposes a critical framework for the future
of AI. It emphasises the need to architect
AI systems that not only align with but also
enhance and sustain human values and
societal norms.
Artificial integrity goes beyond tradi-
tional AI ethics. As AI ethics is the input,
artificial integrity is the outcome advocating
a context-specific application of ethical prin-
ciples, ensuring AI's alignment with local
norms and values.
Underscoring the importance of AI
systems being made socially responsible,
ethically accountable, and inclusive,
especially of neurodiverse perspectives,
the concept represents a deliberate design
approach where AI systems are embedded
with ethical safeguards, ensuring that they
support and enhance human dignity, safety,
and rights. At its heart, this new paradigm
shift aims to apprehend a symbiotic rela-
tionship between AI and humanity, where
technology supports human well-being and
societal progress, redefining the interaction
between human wit and AI's capabilities.
4 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

Introducing the Concept
of ARTIFICIAL INTEGRITY:
THE PATH FOR

THE FUTURE OF AI
by Hamilton Mann
I
n this complex edifice of artificial intelligence that have been established for its functioning, to
progress, the critical challenge for leaders is to be intrinsically capable of prioritising and safe-
architect a future where the interplay between guarding human life and well-being in all aspects
human insight and artificial intelligence doesn’t of its operation.
merely add value but exponentiates it. This is not just about setting ethical standards,
The question is not as simple as whether humans but about the cultivation of an environment where
or AI will prevail, but how their combined forces can AI systems are designed to enable humans to be
create a multiplicative value-added effect, without guided in using, deploying, and developing AI for
compromising or altering core human values but, on the greater interest of us all, thus including the
the contrary, reinforcing them with integrity. planet, in the most appropriate ways.
AI operating systems intentionally designed
and maintained for that purpose would be those Thus, while AI ethics often focuses on universal
that perform with this characteristic. 1 ethical stances, artificial integrity emphasises
adapting them to specific contexts and cultural
settings, recognising that their application can vary
significantly depending on the context.
ARTIFICIAL INTEGRITY IS ABOUT SHAPING This context-specific adaptation of ethical prin-
AND SUSTAINING A SAFE AI SOCIETAL ciples is crucial because it allows for the creation
FRAMEWORK of AI technologies that are not only led by universal
ethics but also culturally competent and respectful
First, external to AI systems themselves, the of important local nuances, thereby sensitive and
concept of artificial integrity embodies a human responsive to local norms, values, and needs,
commitment to establishing guardrails to build enhancing their relevance, effectiveness, and accept-
and sustain a sense of integrity in the deployment ance in diverse cultural landscapes.
of AI technology, ensuring that as AI becomes
more embedded in our lives and work, it supports Differing from AI ethics, which provide the
the human condition rather than undermines it.
More specifically, it refers to the governance
2 external system of moral standards that AI
technologies are expected to follow, concerned with
of AI systems that adhere to a set of principles questions about right or wrong decisions, human
www.europeanbusinessreview.com 5
holistic AI framework. It seeks to create a unified oper-

ational framework where these diverse perspectives
coalesce.
This integrative approach not only enhances the
innovation potential by leveraging diverse expertise but
also ensures more robust, ethically sound, and socially
responsible AI solutions that are better aligned with
complex real-world challenges and stakeholder needs.
Furthermore, while AI ethics recognises the

5 importance of education on ethics, artificial integ-
rity focuses on learning how to de-bias viewpoints to
rights, equitable benefit distribution, and harm preven- embrace 360-degree societal implications, fostering the
tion, artificial integrity is the operational manifestation inclusion of diverse perspectives, especially from neuro-
of those principles. It ensures that AI behaves in a way diverse groups.
that is consistently aligned with those ethical standards. This approach ensures that the development and
This approach not only embeds ethical considera- deployment of AI technologies tap into the large
tions at every level of AI development and deployment spectrum of human neurodiversity and build neuro-re-
but also fosters trust and reliability among users and silience against the distortion of reality.
stakeholders, ensuring that AI systems are not only tech- It empowers AI systems to be more inclusive and
nologically advanced but also driven in a manner that is reflective of the full range of human experiences and
socially responsible and ethically accountable. cognitive styles, leading to more innovative, equitable,
and socially attuned AI solutions.
Unlike AI ethics, which advocates for external stake-
3 holder inputs and considerations in addressing
the societal stakes related to AI deployment, artificial
integrity encompasses a broader spectrum. It involves ARTIFICIAL INTEGRITY IS A DELIBERATE ACT
integrating stakeholders as active participants in a formal OF AI DESIGN TO RESPECT HUMAN SAFETY
and comprehensive operating ecosystem model. AND DIGNITY
This model positions stakeholders at the heart of deci-
sion-making processes, operational efficiency, employee Core to AI systems themselves, the concept of artifi-
engagement, and customer interactions. It ensures that cial integrity implies that AI systems are developed and
organisations can sustain with integrity while being operate in a manner that is not only ethically sound
powered by AI. according to external standards but do so consistently
Such integration is designed not just for compliance over time and across various situations, without devia-
or ethical considerations, but for elevat ing the organisation from their programmed ethical guidelines.
tion’s overall capacity to adapt and thrive in deploying AI It is a deliberate act of design. It suggests a level of
in harmony with societal stakes. self-regulation and intrinsic adherence to ethical codes,
similar to how a person with integrity would act morally,
Moreover, while interdisciplinary approaches are regardless of external pressures or temptations, main-
4 valued in AI ethics, artificial integrity places a
greater emphasis on deep integration across disciplines,
taining a vigilant stance towards risk and harm, ready to
override programmed objectives if they conflict with the
moving beyond a siloed functional approach to a hybrid primacy of human safety.
functional blueprint. It involves a proactive and preemptive approach, where
This blueprint is characterised by the seamless the AI system is not only reactive to ethical dilemmas as
melding of various fields – technology, social sciences, they arise but is equipped with the foresight to prevent
law, business, and more – to create a cohesive and them.

As thought-provoking as it may While traditional AI enrichment of an artificial moral
sound, it is about embedding arti- compass. This approach ensures that
ficial artefacts into AI that will ethics often see ethical AI systems are not only compliant
govern any of its decisions and assessment as a with ethical standards at their incep-
processes, mimicking a form of peripheral exercise that tion but remain dynamically aligned
consciously made actions, while with evolving human values and
ensuring they are always aligned may influence AI design, societal norms over time.
with human values. artificial integrity It represents a significant
This is akin to an “ethical fail- embeds ethical advancement in creating AI systems
safe” that operates under the that are truly responsive and adap-
overarching imperative that no assessment throughout tive to the ethical complexities of
action or decision by the AI system the functioning of the real-world interactions, fostering
should compromise human health, trust and reliability in AI-human
AI’s operating system.
security, or rights. partnerships.
It goes beyond adhering to
ethical guidelines by embedding intelligent safeguards As opposed to AI ethics, which tend to focus on
into its core functionality, ensuring that any potential
harms in the interaction between AI and humans are
7 establishing guidelines for responsible AI design
and usage, artificial integrity, on the other hand, stresses
anticipated and mitigated. the importance of integrating continuous and autonomous
This approach embeds a deep respect for human feedback mechanisms, allowing AI systems to evolve and
dignity, autonomy, and rights within the AI system’s improve in response to real-world experiences, user feed-
core functionality. back, and changing societal norms.
This proactive approach ensures that AI systems
More specifically, while traditional AI ethics often remain relevant and effective in diverse and dynamic
6 see ethical assessment as a peripheral exercise that
may influence AI design, artificial integrity embeds
environments, fostering adaptability and resilience in AI
technologies.
ethical assessment throughout the functioning of the It transcends static compliance, enabling AI to be
AI’s operating system. more attuned to the complexities of human behaviour
This continuous learning and adjustment in inter- and societal changes, thus creating more robust, empa-
action with humans allows for the development and thetic, and contextually aware AI solutions.
While AI ethics focuses on identifying and This expanded focus ensures a deeper understanding of AI
8 addressing risks that correspond to a given
present term, artificial integrity emphasises a more
decisions and actions, enabling users and stakeholders to not only
comprehend AI outputs but also grasp the underlying rationale,
proactive approach in anticipating potential risks in thus fostering greater transparency, trust, and informed deci-
forward-looking scenario perspectives, including sion-making in AI systems.
long-term and systemic risks, before they even
materialise.
This forward-thinking strategy allows organisa- While AI ethics is human-centred,
tions and societies to not only mitigate immediate AI design is about establishing
concerns but also prepare for and adapt to future
guidelines and principles to ensure
challenges, ensuring sustainable and responsible
AI development that aligns with broader societal that AI technologies are developed
goals and ethical frameworks over time. and used in ways that are ethically
sound and beneficial to humanity.
Although AI ethics heavily emphasises data
9 privacy, artificial integrity also stresses the
importance of data integrity, ensuring that data As we transition to a society where AI’s role in society becomes
used by AI systems is accurate, reliable, and more pronounced, the multidisciplinary approach behind artifi-
representative in order to combat misinforma- cial integrity becomes crucial in guiding our future.
tion and manipulation. This approach would ensure that, as AI systems become more
This comprehensive approach not only protects autonomous, their operational essence remains fundamentally
user information but also enhances the overall aligned with the protection and enhancement of human life,
trustworthiness and effectiveness of AI systems, enshrining a harmonious and collaborative future between AI
providing a more solid foundation for deci- and humanity.
sion-making and reducing the risk of errors and
biases that can arise from poor-quality data.
As AI ethics discusses accountability and

10 explainability, artificial integrity broadens
the focal point to include the trade-offs between
ARTIFICIAL INTEGRITY IS A STANCE FOR AI TO SERVE
THE EMPOWERMENT OF HUMANITY
explainability and unexplainability challenges, as Central and, thus, both internal and external to AI systems, the
well as guidelines to fulfil not just explainability concept of artificial integrity embodies an approach where the
but interpretability. relationship between human and AI supports the human condi-
tion rather than undermines it.
The aim is to anchor the role of AI in acting as a partner to
humans, facilitating their work and life in a way that is ethically
aligned and empowering.
It refers to AI integration in society that is designed and
deployed with the intent to augment, rather than replace, human
abilities and decision-making. These AI systems are crafted to
work in tandem with humans, providing support and enhance-
ment in tasks while ensuring that critical decisions remain under
human control.
This is not just about the ethical user-friendliness but about the
fundamental alignment of AI systems with human ethical princi-
ples and societal values. It involves a deep understanding of the
human context in which AI operates, ensuring that these systems

are not only accessible and intuitive but also respectful
of human agency and societal norms.
In essence, while AI ethics is human-centred, AI
design is about establishing guidelines and principles to
ensure that AI technologies are developed and used in
ways that are ethically sound and beneficial to humanity,
and artificial integrity is about creating a harmonious
relationship between humans and AI.
Here, technology is not just a tool for efficiency but
a partner that enhances human life and society in a
manner that is ethically responsible, socially benefi-
cial, and deeply respectful of human values and dignity.
It’s about foreseeing and sustaining a society model
assisted or augmented by AI systems that not only adhere
to ethical norms but also actively contribute to human
well-being, integrating seamlessly with human activities
and societal structures.
This approach is focused on ensuring that AI Each part of this matrix illustrates a distinct narrative about
advancements are aligned with human interests and the future of a human AI-assisted society, presenting us with a
societal progress, fostering a future where AI and strategic imperative: to harmonise the advancement of tech-
humans coexist and collaborate, each playing their nology with the enrichment of human capability and will.
unique and complementary roles in advancing society This is a non-negotiable condition in achieving the sense
and improving the quality of life for all. of integrity rooted in the functioning of AI operating systems
This paradigm shift from mere compliance to proac- for an AI that does not diminish human dignity or potential
tive contribution represents a more holistic, integrated but rather enriches it.
approach to AI, where technology and humanity work
together towards shared goals of progress, well-being, 1. MARGINAL MODE:
and integrity. This part of the matrix reflects scenarios where both human
As we seek to chart a course where the AI of tomorrow intelligence and artificial intelligence have a subdued,
not only excels in its tasks but does so with an underpin- modest, or understated impact on value creation.
ning ethos that champions and elevates human labour,
creativity, and well-being, maintaining and preserving
the equilibrium at the right level, it dares us to question
not only the essence of value but also the vast potential.
This conscientious perspective is especially perti-
nent when considering the impact of AI on society
where the balance between “human value added”
and “AI value added” is one of the most delicate and
consequential.
In navigating this complexity, we must first ensure
not only to delineate the current landscape where
human wit intersects with the prowess of AI but also
serves as a compass guiding us towards future terrains
where the symbiosis of man and machine will redefine
worth, work, and wisdom.
This balance could be drawn through the perspec-
tive of four different modes:
In such a context, we encounter tasks and roles where AI, given the diminishing returns relative to the
neither humans nor AI provide a significant value add. It technology’s implementation cost.
encapsulates a unique category of tasks where the marginal Moreover, the progression of AI technology
gains from both human and artificial intelligence inputs seems to follow a U-shaped pattern of job transfor-
are minimal, suggesting that the task may be either too mation. Early on, automation addresses tasks that
trivial to require significant intelligence or too complex are simple for humans (low-hanging fruit), yet as
for current AI capabilities and certainly not economically AI develops, it starts to tackle more complex tasks,
worth the human effort. potentially leaving behind a trough where tasks are
This mode might typically also involve foundational too trivial for AI to improve upon but also of such
activities where both human and AI roles are still being low value that they do not warrant a significant
defined or are operating at a basic level. human contribution (Brynjolfsson & McAfee, 2014).
It represents areas where The risk in this quadrant
tasks are often routine and On the AI front, although is threefold:
repetitive and do not substan- technologies like optical Firstly, complacency and
tially benefit from advanced obsolescence are the primary
cognitive engagement or AI
character recognition (OCR) risks here.
contributions and may not can digitise documents, they If neither humans nor AI
even require much interven- may struggle with handwritten are adding significant value,
tion or improvement, often it may indicate that the task
remaining straightforward or poorly scanned materials, is outdated or could be at risk
with little need for evolution providing little advantage over of being superseded by more
or sophistication. humans in terms of quality. innovative approaches or
Changes within this area technologies. The task of the
are often small-scale, incremental, or may represent a role might become completely redundant with
state of equilibrium where neither human nor AI contri- the advent of a more sophisticated approach and
butions dominate or are significantly enhanced. processing technologies.
An example is the routine scanning of documents for
archiving. While humans perform these tasks adequately,
the work is monotonous, often leading to disengagement
and errors.
On the AI front, although technologies like optical char-
acter recognition (OCR) can digitise documents, they may
struggle with handwritten or poorly scanned materials,
providing little advantage over humans in terms of quality.
These tasks don’t offer substantial gains in efficiency or
effectiveness when automated, due to their simplicity, and
the return on investment for deploying sophisticated AI
systems may not be justifiable.
This concept aligns with the “task routineness
hypothesis”, which posits that routine tasks are less
likely to benefit from human creativity or AI’s advanced
problem-solving skills (Acemoglu & Autor, 2011).
A study from the McKinsey Global Institute (Manyika
et al., 2017) further elaborates on this by suggesting that
activities involving data collection and processing are often
the most automatable. However, when these tasks are too
simplistic, they might not even justify the investment in

Secondly, for the workforce, these roles are at roles that fall into this low-value quadrant to more engaging
high risk of automation despite the low value added and productive ones that either AI or humans – or a combi-
by AI, because they can often be performed more nation of both – can significantly enhance.
cost-effectively by machines in the long run. Therefore, strategic planning is essential to ensure
A real-world example of this risk materialising that the workforce is prepared for transitions and that the
is in the manufacturing sector, where automation benefits of AI are harnessed without exacerbating socioec-
has been progressively adopted for tasks such as onomic disparities.
assembly line sorting, leading to job displacement.
Research has highlighted this trend and the 2. HUMAN-FIRST MODE:
potential socioeconomic impact, as indicated This side of the quadrant places significant emphasis on
by Acemoglu and Restrepo’s paper, “Robots and the critical roles of human cognition, ethical judgement,
Jobs: Evidence from US Labor Markets” (Journal of and intuitive expertise, with AI taking on a secondary or
Political Economy, 2020), which examines the nega- assistive role.
tive effects of industrial robots on employment and
wages in the US.
Thirdly, from an organisational perspective,
persisting with human labour in such tasks can
lead to a misallocation of human talent, where
employees could instead be upskilled and moved to
roles that offer higher value addition.
The implications of this quadrant for the
labour market are significant, as it often points
to jobs that may be at high risk of obsolescence
or transformation.
There is a growing need for reskilling and
upskilling initiatives to transition workers from
Here, human skills and decision-making are at the

forefront, especially in situations requiring emotional intel-
ligence, complex problem-solving, and moral discernment.
It underscores scenarios where the depth of human
perception, creativity, and interpersonal skills are vital,
where the complexity and subtlety of human cognition
are paramount, and where AI, while useful, currently falls
short and cannot yet replicate the full spectrum of human
intellectual and emotional capacity.
In this sphere, the value derived from human involve-
ment is irreplaceable, with AI tools providing auxiliary
support rather than core functionality.
This is particularly evident in professions such as
healthcare, education, social work, and the arts, where
human empathy, moral judgement, and creative that define these roles maintain their relevance and
insight are irreplaceable and are critical to the importance in the face of technological progress.
value delivered by professionals. The risk in this quadrant comes from misunder-
High-stakes decision-making roles, creative indus- standing the role AI should play in these domains.
tries, and any job requiring deep empathy are areas There is a temptation to overestimate AI’s
where human value addition remains unrivalled. current capabilities and attempt to replace human
For example, in the field of psychiatry, a prac- judgement in areas where it is critical.
titioner’s ability to interpret non-verbal cues, offer An example is the justice system, where AI
emotional support, and exercise judgement based tools are used to assess the risk of recidivism. As
on years of training and experience is paramount. pointed out in the work of Angwin et al. (2016) in
While AI can offer supplementary data analysis, their analysis of the COMPAS recidivism algorithm,
it cannot approach the empathetic and ethical published in Machine Bias by ProPublica, AI can
complexities that humans navigate intuitively. perpetuate biases present in historical data, leading
Empirical research supports this perspective, high- to serious ethical implications.
lighting domains where the human element is crucial. AI systems lack the moral and contextual
For instance, studies on patient care indicate reasoning to weigh outcomes beyond their data
that, while AI can assist with diagnostics and infor- parameters, which could lead to injustices if relied
mation management, the empathetic presence upon excessively.
and decision-making capabilities of a healthcare Therefore, while AI can process and offer insights
provider are central to patient outcomes and satis- based on vast data sets, human beings are para-
faction (Jha & Topol, 2016). mount in applying those insights within the complex
The essential nature of human input in these fabric of social, moral, and psychological contexts.
areas is also supported by studies on job automa- Understanding the boundary of AI’s utility
tion potential, which show that tasks requiring high and the irreplaceable value of human intuition,
levels of social intelligence, creativity, and percep- empathy, and ethical judgement is essential in
tion and manipulation skills are least susceptible to maintaining the integrity of decision-making in
automation (Arntz et al. 2016). these critical sectors.
This is echoed in the arts, where creativity and
originality are subjective and deeply personal, 3. AI-FIRST MODE:
reflecting the human experience in a way that cannot This perspective indicates a technological lean,
be authentically duplicated by AI (Boden, 2009). with AI driving the core operations.
Furthermore, in the context of service indus-
tries, the SERVQUAL model (Parasuraman et al.,
1988) demonstrates that the dimensions of tangi-
bles, reliability, responsiveness, assurance, and
empathy heavily rely on the human factor for
service quality, hence substantiating the need for
human expertise where AI cannot yet suffice.
While AI may offer supplementary functions,
the nuances of human expertise, interaction, and
empathy are deeply entrenched in these high-
value areas.
As such, these sectors are less likely to experi-
ence displacement by AI, instead possibly seeing AI
as a tool that supports human roles.
The continual advancement of AI presents a
moving frontier, yet the innate human attributes

Such an approach is prevalent where the unique patterns, and execute trades at a speed and volume
strengths of AI, such as processing vast amounts of unattainable for human traders.
data with unmatched speed and providing scalable These systems can also be employed in regula-
solutions, take precedence. It often aligns with tasks tory compliance, where they continuously monitor
where the precision and rapidity of AI offer a clear transactions for irregularities much more efficiently
advantage over human capability. than human counterparts
In this domain, AI stands at The acceleration of big-data (Arner et al., 2016).
the forefront of operational execu- The main inherent risks
tion, bringing transformative analytics is one area where in this quadrant are also
efficiency and enhanced capabili- AI demonstrates substantial multifaceted.
ties to activities that benefit from value, as it can uncover First, there is the risk of
its advanced analytical and auton- over-reliance on AI systems,
omous functionalities. insights from data sets too which can lead to complacency
Here, the capabilities of AI are large for human analysts to in oversight. For instance,
leveraged to also perform tasks that in the case of the Flash Crash
process in a timely manner,
generally do not benefit substan- of 2010, rapid trades by algo-
tially from human intervention. as evidenced by research rithmic systems contributed
This AI-first advantage has from Hashem et al. (2015). to a severe and sudden dip in
been extensively documented in stock prices.
the literature, with AI systems outperforming humans in Secondly, while AI can perform these tasks with
data-intensive tasks across various domains. remarkable efficiency, they operate within the
The acceleration of big-data analytics is one area confines of their programming and can sometimes
where AI demonstrates substantial value, as it can miss out on the “bigger picture”, which can only be
uncover insights from data sets too large for human understood in a broader economic, social, and geopo-
analysts to process in a timely manner, as evidenced by litical context.
research from Hashem et al. (2015). Moreover, AI’s dominance in such areas could lead to
An exemplar of this dynamic is evident in areas significant job displacement, raising concerns about the
such as the financial sectors, especially those involved future of employment for those whose jobs are suscep-
in high-frequency trading, where algorithmic trading tible to automation. This shift necessitates a societal
systems can execute transactions based on complex and economic adjustment to manage the transition for
algorithms and vast amounts of market data, recognise displaced workers (Acemoglu & Restrepo, 2020).
Lastly, and especially in this quadrant, ethical considerations

are paramount.
While human input does not significantly enhance these tasks,
the tasks themselves are not devoid of ethical considerations, despite
minimal emotional involvement.
AI systems can perpetuate biases present in their training data,
a concern that has been raised in numerous studies, including by
Barocas and Selbst (2016).
There is the ethical consideration of ensuring that these algo-
rithms operate fairly and transparently, as their decisions can have
wide-reaching impacts on the market and individual livelihoods. The
growing field of explainable AI (XAI) aims to address this, ensuring
that AI’s decision-making processes can be understood by humans,
thereby maintaining a necessary level of trust and accountability in
these high- stakes influential systems.
While AI’s prowess in data processing and routine task automation Here, the focus is on crafting roles and
underscores its high value addition in certain tasks, the importance processes to capitalise on their respective
of human oversight for ethical considerations is a critical aspect that advantages. Human creativity and moral
highlights the need for a collaborative approach between humans reasoning complement AI’s analytical
and AI systems to ensure that ethical standards are maintained. efficiency and pattern recognition.
The interplay of AI’s technical efficiency with human ethical This setting is characteristic of
judgement forms the crux of responsible AI deployment in this quad- forward-thinking workplaces that aim
rant, ensuring that technological advancement involves the careful for a cohesive strategy, maximising the
consideration of the potential impact of AI-assisted decisions on collective benefits derived from both
individuals and society, including the overarching moral implica- human and technological assets.
tions of delegating decisions to machines, so it does not come at the In this environment, the fusion of
cost of ethical integrity. human insight and AI’s precision culmi-
nates in an optimal alliance, propelling
4. FUSION MODE: tasks to new heights of effectiveness.
This area exemplifies a harmonious blend of human intellect and Such a paradigm fosters an atmos-
AI prowess. phere where AI serves as an enhancer
of human skills, ensuring that both
elements are essential to superior
performance and more nuanced
decision-making processes.
This collaboration represents an
ideal in task execution and strategic
planning, offering comprehensive bene-
fits that neither humans nor AI could
achieve in isolation.
Scientific evidence that supports this
synergy comes from various fields.
A study by Rajkomar et al. (2018)
highlights how AI can assist physi-
cians by providing rapid and accurate
diagnostic suggestions based on
machine learning algorithms that

process electronic health records, thus personalise learning experiences in ways that are responsive to
improving patient outcomes. individual student needs, thus supporting educators to tailor their
Such collaboration is particularly teaching strategies effectively.
evident in the realm of medical surgeries. This area underscores a future of work in which AI augments
For example, in image-guided surgery, human expertise, rather than replaces it, fostering a collaborative
AI enhances a surgeon’s ability to differ- paradigm where the complex, creative, and empathetic capacities
entiate between tissues, allowing for of humans are complemented by the efficient, consistent, and high-
more precise incisions and volume processing capabilities of AI.
reduced operative time. In the realm of complex As previously, one of the risks
However, despite the associated with this integration is
clear advantages of AI, the problem-solving and over-reliance on AI, which might
surgeon’s experience and innovation, human creativity lead to complacency.
judgement remain irre- is irreplaceable, even In AI-assisted surgery, a malfunc-
placeable, particularly for tion or misinterpretation of data by
making nuanced decisions though AI can significantly the AI system could lead to serious
when unexpected variables enhance these processes. surgical errors if the human operator
arise during surgery. over-trusts the AI’s capabilities.
In the realm of complex problem- In this integration also lies the risk of a decline in the manual
solving and innovation, human creativity skills of surgeons. Meanwhile, in the event of AI failure or unfore-
is irreplaceable, even though AI can seen situations beyond AI’s current capabilities, the surgeon’s skill
significantly enhancethese processes. becomes paramount.
Evidence has been demonstrated Another risk is the potential for ethical dilemmas, such as the
on how AI can support engineers and decision to rely on AI’s recommendations or strategy when they
designers by offering a vast array of conflict with the surgeon’s clinical judgement.
design options generated through algo- Additionally, there are concerns about liability in cases of malprac-
rithmic processes, which humans can tice when AI is involved. Who is responsible if an AI-augmented
then refine and iterate upon, based procedure goes wrong – the AI developer, the hospital, the surgeon?
on their expertise and creative insight All together, these four modes underscore a future of work in
(Yüksel et al., 2023). which AI augments human expertise fostering a collaborative para-
Lastly, in educational settings, digm where the complex, creative, and empathetic capacities of
research by Holstein et al. (2017) humans are complemented by the efficient, consistent, and high-
provides evidence that AI can volume processing capabilities of AI.
ALGORITHMIC BOOST
5. NAVIGATING TRANSITIONS:
As we migrate from one quadrant to another, we should
aim to bolster, not erode, the distinctive strengths
brought forth by humans and AI alike.
While traditional AI ethics frameworks might not
fully address the need for dynamic and adaptable
governance frameworks that can keep pace with the
transitions in balancing human intelligence and AI
evolution, artificial integrity suggests a more flexible
approach to govern such journeys.
This approach is tailored to responding to the wide
diversity of developments and challenges brought
by the symbiotic trade-offs between human and AI,
offering a more agile and responsive governance
structure that can quickly adapt to new technological
advancements and societal needs, ensuring that AI
evolution is both ethically grounded and harmoniously Algorithmic boost represents scenarios where AI's
integrated with human values and capabilities. role is significantly elevated to augment processes,
When a job evolves from a quadrant of minimal irrespective of the starting or ending point of the
human contribution. This transition focuses on
human and AI value to one where both are instrumental, harnessing AI either to take the lead in processes
such a shift should be marked by a thorough contempla- where human input is low or to amplify outcomes in
tion of its repercussions, a quest for equilibrium, and an scenarios where the human value is already high.
adherence to universal human values.
For instance, a move away from a quadrant character-
HUMANISTIC REINFORCEMENT
ised by AI dominance with minimal human contribution
should not spell a retreat from technology but a recalibra-
tion of the symbiosis between humans and AI.
Here, artificial integrity calls for an evaluation of
AI’s role beyond operational efficiency and considers its
capacity to complement, rather than replace, the complex
expertise that embodies professional distinction.
Conversely, when we consider a transition toward
less engagement from both humans and AI, artifi-
cial integrity challenges us to consider the strategic
implications carefully. It urges us to contemplate the
importance of human oversight in mitigating ethical
blind spots that AI alone may overlook. It advocates
ensuring that this shift does not signify a regression
but a strategic realignment toward greater value and
ethical integrity.
Different types of transitions or shifts occur as
organisations and processes adapt and evolve in
response to the changing capabilities and roles of Humanistic reinforcement counters the first by
humans and AI. emphasising transitions that increase the human value
added in the equation. This set of transitions may
These transitions are grouped into three main
involve reducing AI's role to elevate human interaction,
types: algorithmic boost, humanistic reinforcement, creativity, and decision-making, thereby reinforcing
and algorithmic recalibration. the human element in the technological synergy.

ALGORITHMIC RECALIBRATION
constantly evaluated against the imperative to nurture and

promote human dignity, creativity, and moral frameworks.
In this age of swift technological advancement, the philosophy
of artificial integrity provides a guiding light, ensuring that our
navigation through the AI-powered matrix of the world not only
celebrates the synergy of human and machine but also protects
the human ethos at the heart of true innovation.
In introducing artificial integrity to the discourse, we set out
to explore the potential transformation of tasks, jobs, and the
collective workforce across industries and, importantly, how the
confluence of AI and human destiny can be guided with vision,
accountability, and a deep-seated dedication to the values that
are quintessentially human.
REFERENCES:
• Acemoglu, D., & Autor, D. (2011). "Skills, tasks and technologies: Implications for
employment and earnings", Handbook of Labor Economics.
Lastly, algorithmic recalibration consists of
• Manyika, J., et al., (2017). "A future that works: Automation, employment, and
transitions that involve a reassessment and productivity", McKinsey Global Institute.
subsequent adjustment of the balance between • Brynjolfsson, E., & McAfee, A. (2014). "The second machine age: Work, progress, and
human and AI contributions. This might mean a prosperity in a time of brilliant technologies", W.W. Norton & Company.
• Acemoglu, D. & Restrepo, P. (2020). "Robots and Jobs: Evidence from US Labor
reduction in AI's role to correct over-automation or
Markets", The University of Chicago Press Journals.
a decrease in human input to optimise efficiency • Jha, S., Topol, E.J., (2016). "Adapting to Artificial Intelligence: Radiologists and
and capitalise on advanced AI capabilities. Pathologists as Information Specialists", JAMA.
• Arntz, M., Gregory, T., Zierahn, U., (2016). "The Risk of Automation for Jobs in
OECD Countries", OECD Social, Employment and Migration Working Papers.
• Boden, M.A., (2009). "Computer models of creativity", AI Magazine.
Together, these sets of transitions provide a • Parasuraman, A., Zeithaml, V.A., & Berry, L.L. (1988), SERVQUAL model
comprehensive framework for understanding • Angwin et al. (2016), "How We Analyzed the COMPAS Recidivism Algorithm", "Machine
Bias" by ProPublica.
and strategising the future of work, the role • Hashem, I.A.T., Yaqoob, I., Anuar, N.B., Mokhtar, S., Gani, A., & Khan, S.U. (2015).
of AI, and the optimal collaboration between "The rise of 'big data' on cloud computing: Review and open research issues",
Information Systems.
human intelligence and artificial counterparts. • Arner, D.W., Barberis, J.N., & Buckley, R.P. (2016). "The evolution of fintech: A new
They reflect an ongoing dialogue that focuses post-crisis paradigm?", SSRN Electronic Journal.
• Acemoglu, D., & Restrepo, P. (2020). "Robots and jobs: Evidence from US labor
not only on enhancing human skills and markets", Journal of Political Economy.
leveraging advanced technology but also on • Barocas, S., & Selbst, A.D. (2016). "Big Data's Disparate Impact", California Law Review.
• Rajkomar, A., Dean, J., & Kohane, I. (2018). "Machine Learning in Medicine", The New
maintaining artificial integrity. England Journal of Medicine.
• Yüksel, N., Börklü, H.R., Sezer, H.K., & Canyurt, O.E. (2023). "Review of artificial
This ensures that, as we find the right balance intelligence applications in engineering design perspective", Engineering Applications
between the two, we do so with a commitment to of Artificial Intelligence.
• Holstein, K., McLaren, B.M., & Aleven, V. (2017). "Intelligent tutors as teachers' aides:
integrity’s standards, ensuring that AI systems are Exploring teacher needs for real-time analytics in blended classrooms", The Seventh
transparent, fair, and accountable. International Learning Analytics & Knowledge Conference.
Upholding artificial integrity is paramount, as

it governs the trustworthiness of AI and secures ABOUT THE AUTHOR
its role as a beneficial augmentation to human
capacity rather than a disruptive force. Thus, the Hamilton Mann is the Group VP for Digital
journey towards technological advancement and Marketing and Digital Transformation
automation is navigated with a conscientious effort at Thales. He is also a Guest Lecturer
to sustain both innovation and human values. at INSEAD and a Senior Lecturer at
Artificial integrity becomes a compass by which HEC Paris and EDHEC Business School.
we can steer through this evolving landscape. Additionally, he serves as a mentor at the
It beckons us to maintain a careful balance, MIT Priscilla King Gray (PKG) Center and hosts The Hamilton
where the integration of AI into our tasks is Mann Conversation (www.hamiltonmannconversation.com).
THE PATH
TOWARDS
TRUSTWORTHY
AI IS NO TECH
BUT A HUMAN
INTELLIGENCE
TEST
By Hamilton Mann
In the frenzy to champion the

potential of trustworthy AI, the
recent moves from tech giants offer
a reflective pause about one of the
most, if not the most important
aspects of AI, which, paradoxically,
is seldom discussed: the challenge it
poses to human intelligence.
THE QUEST FOR AI’S SENSORY As we analyse OpenAI and Meta's innovations,
PERCEPTION the growing capability of AI to emulate human-like
behaviour cannot be ignored.
With OpenAI's ChatGPT flaunting sensory capabilities However, a closer look, underpinned by scientific
and Meta introducing AI chatbot personalities, the evidence, unveils the intricate layers involved and
crescendo of AI's role in our lives is unmistakable. prompts important inquiries about the direction AI
While these advancements showcase the leaps AI should take.
has made, there's a subtext here: these AI systems To begin with, the architecture of many AI
are mirroring complex human communication models is inspired by human neural networks. For
capabilities. instance, the deep learning models use layers of
It's easy to get entangled in the glitz of AI's interconnected nodes, reminiscent of how neurons
capabilities and miss the fundamental question: are connected in the human brain. A research paper
should AI aim to mirror human faculties or should it from Angela D. Friederici titled “The Brain Basis of
charter a different course? Language Processing: From Structure to Function”

published in the American Psychological Society Lastly, through the prism of ethical and
Journal in 2011, indicates that when humans engage philosophical lenses, the quest to replicate human
in complex communication, multiple regions of the faculties in AI brings forth ethical dilemmas. The
brain, including Broca’s and Wernicke’s areas, work Human Brain Project (HBP) funded by the European
synchronously. Union seeks to understand the intricacies of the
Similarly, AI models, such as OpenAI's GPT series, human brain, potentially offering insights into
employ multiple layers to generate and interpret text, creating more human-like AIs. But this brings up a
mimicking this orchestrated brain activity. philosophical question: Just because we can replicate
When it comes to grasping semantics, while AI has certain aspects of human cognition in machines, does
made strides in producing human-like text, there's a it mean we should?
distinction between generating syntactically correct While evaluating AI's character may seem akin to
sentences and truly understanding semantics. The understanding human nature, it's crucial to realise
Neurocognition of Language, a book published in 2000 that AI doesn't have personal experiences, emotions,
by Oxford University Press, highlighted that human or consciousness. Instead of anthropomorphising AI,
brains process words and context in tandem, allowing we should aim to understand its unique nature.
for a deeper understanding of language nuances. AI, in As we push for greater intelligence in machines,
contrast, relies heavily on patterns in data without truly it becomes equally crucial to instill boundaries that
grasping the underlying meaning. This distinction guide this intelligence in a responsible manner.
underscores the difference between This won’t be and shouldn’t be
superficial emulation and genuine made by the machine for itself.
comprehension.
While AI can
Diving into emotional intelli- recognise emotional
gence, Meta’s advancements in AI cues, it doesn't
highlight its ability to interpret experience emotions THE COMPLEX
and simulate human ones through AI GUARDRAILS
facial recognition and text analysis.
in the human EQUILIBRIUM
However, scientific studies, such sense, indicating a
as those by Antonio R. Damasio in fundamental disparity Leading voices emphasise the
his book Descartes' Error published importance of guardrails to avoid
between recognition
in 1994, emphasise the intrinsic AI's pitfalls. Yet, historically, revo-
link between emotions and human and experience. lutionary technology faced similar
consciousness. While AI can recog- trepidations. Cars, when first intro-
nise emotional cues, it doesn't experience emotions in duced, faced skepticism, with critics demanding
the human sense, indicating a fundamental disparity speed-limiting devices to ensure safety. Imagine
between recognition and experience. limiting vehicles to a pedestrian's pace! In a bid to
On the artistic spectrum, AI models, such as contain AI, are we stifling its potential?
DALL·E by OpenAI, can generate creative images, but The introduction of electricity transformed homes
their “creativity” is constrained by patterns in their and industries but also brought risks such as electrical
training data. The research paper “DALL·E: Creating fires and electrocution. As infrastructure and
Images from Text” published in the OpenAI Blog in regulations evolved, safety improved without curbing
2021 highlighted that while AI can mimic certain the transformative power of electricity. The core
creative processes, it lacks the intrinsic spontaneity principle here is adaptability. As society understands
and serendipity inherent in human creativity. Its the potential dangers of a particular technology,
creativity, unlike that of humans, isn't influenced by a guidelines can be adjusted to ensure safety without
lifetime of diverse experiences, emotions, or moments inhibiting innovation.
of serendipity. Instead, it relies on vast quantities of A look back at technological milestones can offer
data and learned patterns. instructive parallels.
As society
understands the
potential dangers of a
particular technology,
guidelines can be
adjusted to ensure
safety without
inhibiting innovation.
Historically, the aviation industry underwent the technology. It ensures that as technology advances,
multiple safety iterations before reaching today's its safety and ethical implications are addressed in
standards. Early planes faced numerous accidents, real time, striking a balance between potential and
leading to skepticism about commercial flight. precaution.
However, over time, rigorous testing, improved While it's valid to raise concerns about stifling the
design, and advanced regulations have made flying potential of AI with excessive guardrails, appropriately
one of the safest modes of transportation. Iterative calibrated precautions can, in fact, bolster innovation
improvement based on accumulated data and real- by building trust and ensuring broad societal
world experiences can refine both technology and its acceptance.
safety protocols. Rather than stifling potential, these Finding the right equilibrium is as essential as
refinements can bolster public trust and facilitate understanding the moral principles that shape these
broader adoption. boundaries, giving AI its ethical foundation.
Similarly, the development of nuclear energy saw Again, the machine won't and shouldn’t autono-
significant hesitancy, given the catastrophic potential mously generate this for itself.
of mishaps. However, meticulous regulations, safety
protocols, and international pacts have allowed
nations to harness nuclear power without widespread
disasters. Properly calibrated regulations can serve THE AI MORAL COMPASS
dual purposes: ensuring public safety and providing
a structured framework within which innovations Anthropic and Google DeepMind's attempts to create
can flourish. Overly strict regulations might stifle AI constitutions—core principles guiding AI behaviour
potential, but a complete lack can result in distrust —are commendable. However, once the authority of
and potential misuse. certain final principles is established, other avenues
Conversely, the Internet's rise was swift, of understanding are often dismissed. By framing
catching many regulators unprepared. While it AI's potential within our current ethical constructs,
has democratised information, the lack of initial we might inadvertently limit its vast potential. The
guardrails has led to issues such as cyberbullying, creation of an AI constitution should be evolutionary,
misinformation, and data privacy concerns, and those rather than prescriptive.
are still a primary concern today. The challenge has From a historical perspective, Thomas S. Kuhn,
been retroactively implementing guidelines without in his influential book “The Structure of Scientific
curtailing the web's intrinsic freedom. Revolutions” published by the University of Chicago
Rapidly evolving technologies can benefit from Press in 1962, posited that science progresses
early, flexible guardrails that evolve in tandem with through paradigms—widely accepted frameworks

of understanding. However, once a paradigm takes guiding principles will be questioned in tandem.
hold, it often constrains alternative viewpoints That said, with a clear moral foundation set for AI,
and approaches. This can be applied to AI ethics: a we must then ensure that it truly represents everyone,
too-rigid AI constitution might become the de facto emphasising the importance of inclusiveness.
paradigm, constraining alternative ethical approaches Yet again, the machine won't and shouldn’t create
and potentially stifling innovation. this on its own.
Turning to economics, behavioural economists
like Herbert A. Simon have argued that humans
often make decisions based on “bounded rationality”,
limited by the information they have, cognitive THE PATH TOWARDS AI
limitations, and the finite amount of time to make a INCLUSIVENESS
decision. If AI is constrained strictly by our current
bounded understanding of ethics, it may not explore Reinforcement Learning by Human Feedback (RLHF)
potentially better solutions outside of these bounds. as a method to refine responses generated by AI has
Delving into psychology, research from the field faced criticism for being primitive.
of moral psychology, such as Jonathan D. Haidt's But let's examine this.
work on moral foundations theory suggests that If AI learns from human feedback, doesn’t it
human morality is complex, multidimensional, and reflect our collective psyche? Instead of overhauling
varies across cultures. If we overly standardise an AI this method, diversifying the pool of evaluators might
constitution, we may overlook or offer richer feedback, reflecting a
undermine this richness, leading tapestry of human perspectives.
to AI systems that don't account for Critically, multiple Critically, multiple studies have
the vast tapestry of human values. studies have shown shown that AI models can inherit and
Drawing from natural processes that AI models can amplify human biases, especially
in evolutionary biology, nature's if they are trained on biased data.
diversification strategy ensures
inherit and amplify For example, “Semantics derived
survival and adaptation. Species human biases, automatically from language corpora
that were too specialised and especially if they are contain human-like biases”, a 2018
inflexible often went extinct when trained on biased data. Princeton research study published
conditions changed. Similarly, an in the journal Science demonstrated
AI that is too narrowly confined that commercial algorithms showed
by a rigid set of principles may not adapt well to gender biases by associating male words with careers
unforeseen challenges. and female words with family. This suggests that if the
Exploring genetic frontiers in the realm of human feedback in RLHF comes from a homogenous
bioethics, the introduction of Clustered Regularly group, the resultant AI behaviour might also be skewed.
Interspaced Short Palindromic Repeats (CRISPR) – a
technology that research scientists use to selectively
modify DNA, adapted for laboratory use from a
naturally occurring defense mechanism in bacteria
that allows them to recognise and destroy foreign DNA
from viruses – has sparked debates about the limits
of genetic modification. Some argue for restraint
based on current ethical principles, while others
believe there's a need for evolving ethical guidelines
as we learn more about the technology. This can serve
as an analogy for AI: as we discover more about its
capabilities and implications, the evolution of our
From a global standpoint, cross- Moving towards inhibition, and short-term memory,
cultural psychology has uncovered collectively termed “executive
significant differences in how moral diversity is a key control”. Incorporating neurodiverse
values are prioritised in different condition for evaluators in RLHF can provide
cultures. For instance, a study titled “Is It developing AI's varied cognitive feedback, leading to
Good to Cooperate? Testing the Theory of a more robust AI model.
Morality-as-Cooperation in 60 Societies”
inclusiveness. From a slightly different angle
by Oliver S. Curry et al., published in It's an essential but reaching a similar conclusion,
the University of Chicago Press Journals precursor for James Surowiecki's book, The Wisdom
in 2019, found that while certain moral recognising and of Crowds, presents evidence that
values were universally recognised, their collective decisions made by a diverse
interpretation varied across cultures.
actively countering group often lead to better outcomes
Thus, a diverse pool of evaluators in biases, ensuring than even the best individual decision.
RLHF can offer a more holistic view of AI's consistent When applied to RLHF, this suggests
what's “right” or “acceptable”.
On the neural front, neuroscientific
utility and fairness. that a diverse group of evaluators can
provide more accurate and balanced
research indicates that people from feedback than a select few experts.
different backgrounds or with different neurological Reflecting on past shortcomings, there have been
makeups process information differently. For instances where a lack of diversity in evaluators has
example, studies have shown that bilingual individuals led to unintended AI behaviour. For example, the
can process certain language tasks differently from racial and gender bias in certain facial recognition
monolinguals. One of the renowned experts in systems can be traced back to a lack of diversity in
this field is Dr. Ellen Bialystok, who has conducted training data.
numerous studies on bilingualism and its effects on Failures in using the RLHF method to ensure
cognitive processes. For instance, Bialystok's research improved response of AI are not because of the
study titled “Bilingualism: Consequences for Mind method itself, but due to the lack of diversity. Ensuring
and Brain” published in Trends in Cognitive Sciences a diverse pool for RLHF can help mitigate such pitfalls.
in 2012 has shown that bilinguals often outperform Moving towards diversity is a key condition
monolinguals in tasks that require attention, for developing AI's inclusiveness. It's an essential
precursor for recognising and actively countering
biases, ensuring AI's consistent utility and fairness.
Similarly here, this won't and shouldn’t be self-
managed by the machine.
THE BATTLE AGAINST AI’S INHERENT

BIASES
Red-teaming, a process of “breaking” AI to understand

its vulnerabilities, while robust, resembles the older
software testing methods. By focusing on adversarial
testing, we might be swayed by collective consensus
rather than individual merit.
While red-teaming aims to find vulnerabilities in AI
systems by simulating adversarial attacks, the nature
of these attacks often reflects known vulnerabilities.

potential risks. This underscores the need for a
globally diverse team for comprehensive red-teaming.
While it's crucial to understand AI's response in
worst-case scenarios, an overemphasis can lead to
neglect of more mundane but equally critical issues.
A case in point is Microsoft's Tay, an AI chatbot that
began tweeting inappropriate content not due to
an adversarial attack but because of the data it was
exposed to. A strict red-teaming approach might
miss such vulnerabilities that arise from the model's
“Towards Evaluating the Robustness of Neural regular interactions.
Networks”, a research study by Nicholas Carlin and AI models, especially those incorporating some
David A. Wagner from UC Berkeley published in form of online learning, evolve over time. A one-time
arXiv in 2017, highlighted that adversarial examples red-teaming might not be enough. There is a need
(perturbed inputs designed to fool machine learning for continuous and dynamic testing methodologies
models) in one domain can be starkly different from tailored for AI, as models can drift from their initial
another. By concentrating only on known issues, we behaviour due to continuous updates.
might neglect emergent risks specific to the evolving Ultimately, addressing biases is an ongoing
nature of AI. process, pushing the boundaries and goals of AI to
In addition, traditional software red-teaming continually adapt and evolve.
often focuses on a limited set of potential threats or As with before, it's not and it shouldn’t be the
vulnerabilities. However, the complexity of modern machine's design to craft this on its own.
AI models, like deep neural networks, demands
an extensive landscape of possible threats. A 2018
paper titled “Adversarial Risk and the Dangers of
Evaluating Against Weak Attacks” published in arXiv THE AI'S EVOLVING FINISH LINE
by Jonathan Uesato et al., demonstrated that larger
neural networks, while more accurate, are often more We often perceive AI as a problem awaiting a solution,
susceptible to adversarial attacks, implying a vast yet we must not forget the rich tapestry of human
attack surface. experiences. The goal for AI's future should not solely
Moreover, human biases can infiltrate the be to forge an infallible model but to consider how we
red-teaming process. “Gender Shades: Intersectional might embrace its inherent imperfections, just as we
Accuracy Disparities in Commercial Gender do with humanity.
Classification” the study co-authored by Joy As we stand poised at the intersection of AI's
Buolamwini and Timnit Gebru and published in rapid advancement and its profound implications for
Proceedings of Machine Learning Research in 2018 humanity, our endeavour should be to co-create an
highlighted that AI systems trained in one cultural AI ecosystem that mirrors the finest of human ideals,
context might exhibit vulnerabilities that are entirely convictions, and hopes. In doing so, we must always
overlooked by red-teamers from that same context, remember that no machine can, or should, ever
simply because their own biases blind them to supplant human critical thought.
ABOUT THE AUTHOR
Hamilton Mann is the Group VP for Digital Marketing and Digital Transformation at Thales. He is
also a Guest Lecturer at INSEAD and a Senior Lecturer at HEC Paris and EDHEC Business School.
Additionally, he serves as a mentor at the MIT Priscilla King Gray (PKG) Center and hosts The
Hamilton Mann Conversation (www.hamiltonmannconversation.com).
ENHANCING OPERATING MODELS'

QUOTIENT(AIQ)
by Hamilton Mann
W
ith the weekly drumbeat of Generative AI advancements
As organisations tap into the power and corporate leaders signalling the need for their organ-
of Generative Artificial Intelligence isations to make progress in harnessing the power of AI,
to improve business outcomes, it is larger questions are emerging for these same executives to address.
imperative to examine how it affects their In addition to the ethical challenges that AI presents for their
operating models. Only by doing this will customers, employees, and society, companies must grapple with
companies leverage the power of AI while how AI will fundamentally shift their operating model including the
avoiding its pitfalls. workforce they employ today.

Ignoring the seismic shifts brought about while remaining flexible enough to navigate the
by AI, and in particular, Large Language 20% of unforeseen challenges.
Models (LLMs) is no longer a viable option AI can lend unprecedented proficiency in
for organisations. managing these unpredictable elements, simulta-
The torrent rise of AI, championed by neously elevating the performance within that 80%
industry titans such as OpenAI, Google, Meta, of predictable events.
Microsoft, and Nvidia, is rapidly reshaping The primary concern lies in establishing a
how work gets done and how companies dynamic organisational structure that opti-
operate and deliver value to their mises efficiency in addressing these
customers and shareholders. regular tasks, allowing AI to enhance
Let's explore the key underlying Preparing an this productivity while maximising
components of organisations' oper- organisation for AI agility in responding to the organisa-
ating models—namely, organisational tion's blind spots.
structure, people, processes, tech- is less a matter of Let's not merely ask how to intro-
nology, and culture—while shaping the stringent modification duce AI into organisations. Instead,
informal and often unwritten mecha- and more a journey let's question how to transform
nisms that capture their essence and organisations and reshape an under-
are most profoundly transformed by towards a fluid standing of what's possible with AI.
AI's influence. organisation
CULTURE: MINDSET OVER

ORGANISATION: BLUEPRINT OVER SKILLSET
STRUCTURE
The digital realm may seem daunting, requiring
In an era dominated by rapid AI advancements, a profound understanding of data, technology,
it's crucial to assess the impact on organi- algorithms, and AI, but this is a misperception.
sations from the holistic perspective of an Amid the surge of digital technology evolu-
organisational blueprint, rather than merely an tion, leaders need to take steps towards dispelling
organisational structure. the myth of digital omniscience, emphasising
The forward-looking and comprehensive a more critical and discerning approach to
nature of a blueprint, designed for adapt- digital understanding.
ability, offers a more inclusive approach that Rather than cultivating an army of data
anticipates future changes and seamlessly scientists and programmers, the focus should be
integrates AI's transformative potential into on fostering a mindset that embraces the poten-
the very fabric of an organisation's operations tial of these systems.
and strategy.
Preparing an organisation for AI is less a
matter of stringent modification and more a
journey towards a fluid organisation. A rigid
adaptation approach tends to breed compla-
cency and a clinging to the status quo in a
defensive posture, often arising from an innate
Photo credit: Unsplash.com
human need for stability.

Conversely, fluid organisation involves
crafting a target model that provides a sense of
reliability in handling 80% of predictable events,
The transition from traditional processes

to digital ones is a journey of exploration—
embracing change, questioning the status quo,
and learning to consider the implications of AI
—all while understanding that the goal is not
mastery, but fluency in concepts such as system
architecture, AI agents, cybersecurity, and data-
driven experimentation.
Moreover, it's imperative to acknowledge
that AI, no matter how advanced, is more than
just a tool.
Unlike other technological tools, AI has
the capacity to learn, adapt, and even make Photo credit: Freepik.com
decisions based on the data it receives. When
referring to AI merely as a ''tool'', it's essential to
ensure its profound implications are not under-
estimated, given the unique form of intelligence
it embodies.
Such a mindset could lead to significant over-
sights, thereby running the risk of undesirable
consequences by overlooking the
necessary precautions required for The transition from Leaders should, therefore, instill
its deployment. traditional processes to a culture of continuous learning,
Blind reliance on them, or flexibility, and adaptability.
reliance without a clear sense of digital ones is a journey Embracing digital technology
purpose or ethical considerations, of exploration— also means acknowledging its
must be the pitfall to avoid. embracing change, impermanence and the need to
True leadership in the digital be proactive rather than reactive.
age isn't about tech prowess but questioning the status Hence, this is not just about under-
the ability to integrate technology quo, and learning standing the current AI capability
meaningfully into broader objec- but anticipating the ones yet to come,
to consider the
tives, ensuring it aligns with human ensuring not just keeping pace with
values and societal positive impact. implications of AI. what is available but also taking steps
Leaders must ensure that the ahead, shaping its very trajectory.
organisation pushes itself and is constantly
challenged intrinsically by its mode of oper-
ation. AI must be approached not as an
infallible oracle but as a powerful ally that, PROCESS: DATA OVER
when used with discernment, can amplify PROCEDURE
human capacities.
Lastly, it's essential to understand that the In the realm of AI, data stands as a center-
very essence of any digital technology is its piece, evolving platforms, tools, and systems,
evolutionary nature. What may be a ground- facilitating greater efficiency and improved
breaking innovation today could become service delivery.
obsolete tomorrow. Relying solely on the tech- As AI begins to take on an increasingly domi-
nical know-how of the present might lead to the nant role in decision-making, a critical challenge
trap of short-sightedness. has emerged: understanding the labyrinthine

data-driven process of AI’s reasoning for the While it's beneficial to involve diverse
sake of trustworthiness. teams in data collection and processing,
Let's move beyond the perfectionism of overemphasis on representation might lead
causality that leads to linear and procedure- to enforced uniformity, suppressing the rich
thinking to embracing the prag- and natural variations in human expression
matism of effectuality. AI must be and experience. Instead, a more balanced
Embracing practices like high- approached not approach would allow AI models to learn
lighting relevant data sections and adapt from the organic nature of data,
contributing to AI outputs or as an infallible including biases, to respond more genuinely
building models that are more oracle but as a to different perspectives.
interpretable could enhance AI powerful ally Finally, let’s rethink the trade-offs of large
transparency. But is transpar- datasets vs small datasets.
ency the only antidote to the trust that, when used The pursuit of larger AI systems by tech
issues with AI? Or could there be with discernment, companies is not merely a race towards
a different approach that not only volume. Larger datasets encompass broader
can amplify
explains AI's decisions but also knowledge, mirroring the vast spectrum of
anticipates its consequences? human capacities. human perspectives.
Leaders must come to terms Reducing the size of a model for the sake of
with the uncomfortable truth that AI’s better understanding might, in fact, diminish
decision-making capabilities often far exceed the depth and richness of insights it can provide.
human comprehension. However, AI can help No matter how meticulously AI is devel-
leaders understand in detail the associated oped and documented, it can never fully
effects of AI’s decision-making capabilities.
For instance, AI could simulate various
scenarios to illustrate the potential outcomes of
its recommendations. This way, AI can be used
to understand the breadth and depth of its own
impact. In a medical setting, AI might recom-
mend a certain treatment plan. Anticipating
the consequences means understanding how
this treatment could affect the patient's health
outcomes, taking into consideration the individ-
ual's unique medical history and circumstances.
It is also worth mentioning that AI's effec-
tiveness is heavily influenced by the data it
processes, forget about an unbiased dataset as
the magic bullet to address biases. There will
always be biases in datasets, as data are orig-
inally produced by humans and the process of
refining them involves humans again.
Embracing the intrinsic nature of bias in data-
sets is a challenge that can lead to more accurate
Photo credits : Unsplash.com
and adaptable AI models. This is achieved by

recognising that total neutrality is a myth and
integrating a diverse range of data to ensure AI
models can respond to various contexts.
It's not only about who curates the data.
grasp the depth of human experiences and The emergence of AI necessitates new skill
biases, leading to inadvertent harm. Hence, sets and competencies, redefining what exper-
the strategy shouldn't be to eradicate bias but tise is essential for delivering value in this new
to acknowledge and manage it, reducing the “AIconomic” era.
risks that are associated while enabling us to But it goes beyond that.
navigate complex human biases and The prospect of AI triggering
patterns effectively. mass unemployment is often
AI will only be truly powerful Embracing the intrinsic overshadowed by optimistic
when it can navigate the complex, nature of bias in datasets predictions based on historical
bias-ridden real world. That will technological revolutions. It is
only be achieved while appreci- is a challenge that can imperative, however, to examine
ating the multifaceted nature of lead to more accurate AI's impact not through the lens
data and developing AI models and adaptable AI models. of the past, but in the context of
that can recognise and adapt to its unique capabilities.
these complexities that are in essence not For instance, the transition from horse-
always “procedurable”. and-buggy to automobiles indeed reshaped
job markets but it did not render human skills
redundant. AI, on the other hand, has the poten-
tial to do just that.
PEOPLE: HUMAN CAPITAL Contrary to the belief that AI should not
VALUE TRANSITIONING OVER create meaningful work products without
RESKILLING human oversight, the use of AI in tasks like
document generation can result in increased
Leaders need to face the new or exacerbated efficiency. Of course, human oversight is
Human Capital challenge AI poses. important to ensure quality, but relegating AI
ORGANIZATION: “BLUEPRINT OVER STRUCTURE”

Distributed Teams
Fluid Organization
AI Augmented Strategy
AI integration Roadmap
Uncertainty Management through AI
Intelligent Organizational Responsiveness
CULTURE: “MINDSET OVER SKILLSET” PEOPLE: “HUMAN CAPITAL OVER RESKILLING”

Digital Fluency Human Capital Value Transitioning
Human-Centric Tech Leadership Transitional Support Policies
AI Stewardship Human-AI Collaboration
Societal AI Alignment Social Capital Safeguarding
Oracle Fallacy Avoidance Human-Centric AI Governance
Strategic AI Questioning AI QUOTIENT Inclusive “AIconomics” standards
TO PIONEERING
AI-ASSISTED
PROCESS: “DATA OVER PROCEDURE” OPERATING TECH: “ETHICS OVER VALUE PROPOSITION”
Multifaceted Data Literacy Collapsed Tech Stack
MODELS Sustainable Development-friendly Infrastructure
AI-Enabled Empathy
Ethical Data Stewardship Standards
Dynamic Bias Management Responsive Ethical AI Framework
AI Augmented Decision-making Trust-Centric AI design
Data-driven Effectual thinking Real-World AI Readiness
AI Safety Nets
+DPLOWRQ0DQQ

It is imperative to also craft comprehen-
sive social and economic policies that provide
immediate relief and long-term support to those
displaced by AI's advancement.
Unemployment benefits, for instance,
could be reevaluated and expanded to cater to
AI-induced job losses.
Moreover, addressing AI displacement
should not solely focus on financial security.
The social and psychological impacts of job loss
— including the loss of identity, self-esteem, and
social networks — are equally significant and
need to be factored into policy planning.
Social support services and career counsel-
ling could be made widely accessible to help
individuals navigate the transition period.
A Human Capital Value Transitioning Analysis
Photo credit: Unsplash.com
can effectively cushion the impact of AI-induced

displacement and build a resilient and inclusive
organisation from AI advancements while safe-
guarding its human capital.
to merely auxiliary roles might prevent us from TECHNOLOGY: ETHICAL STANDS

fully realising its potential. OVER VALUE PROPOSITION
Take Collective[i]'s AI system for instance.
Yes, it may free up salespeople to focus on rela- AI introduces novel policies and standard needs,
tionship building and actual selling, but it could necessitating a reevaluation of decision-making
also lead to a reduced need for human personnel, protocols and organisational conduct.
as AI handles an increasingly larger share of But let’s not think that AI regulation will be
sales tasks. The efficiencies of AI could easily enough to regulate AI.
shift from job enhancement to job replacement, The agile nature of AI evolution has outpaced
creating a precarious future for many roles. the regulation meant to keep it in check. The
Similarly, while OpenAI's Codex may make burden of ensuring that AI tools are used
programming more efficient, it could, in the long ethically and safely thus rests heavily on the
run, undermine the value of human program- shoulders of the companies employing them.
mers. As AI progresses, the line between "basic The role of AI ethics watchdogs and regu-
purposes" and more complex tasks will blur. lation is crucial, but their effectiveness can
Certainly, investments in education and be limited by the rapidly changing land-
upskilling form a key part of any strategy to cope scape of AI. Overly relying on the arrival of
with job displacement due to the rise of AI. This external checks and balances, or acting as if
includes fostering new-age skills that enable waiting for them to first take a stance before
workers to adapt to the changing employment taking action, could lead to complacency
landscape and thrive in AI-dominated sectors. within organisations.
However, this approach alone may not It is thus essential for leaders to foster a
be sufficient. culture of ethical AI development and usage,
Photo credit: Freepik.com

and not just depend on external watchdogs
or regulations.
While government regulations are evolving
to address AI, organisations should proactively
ensure their AI applications are responsible, fair,
and ethical.
It's not just about reaping the benefits of AI
but also about responsibly integrating these instead of just reskilling, and elevating ethical
technologies without causing harm to stake- stances above traditional value propositions.
holders. This necessitates not only technological To navigate this multi-dimensional transfor-
sophistication but also ethical mindfulness and mation effectively, organisations would benefit
societal understanding. from a structured approach to assess their
The example of Zoom — the readiness and progress in
popular video conferencing Addressing AI displacement developing their Operating
software — which recently made should not solely focus on Model’s ‘‘AI Quotient’’.
headlines, raising concerns Companies that can quickly
about an update to their terms-
financial security. The social evolve towards these dimen-
of-service that allows the and psychological impacts of sions will begin to separate
company to use customer data job loss — including the loss themselves from the pack in
to train its artificial intelligence, their respective industries in
illustrates this.
of identity, self-esteem, and terms of the speed and impact
The path to responsible AI social networks — are equally of AI-driven innovations.
deployment is less about waiting significant and need to be AI is like no other tech
for appropriate regulations and wave in history with the poten-
more about fostering a deep
factored into policy planning. tial to empower employees,
understanding and ethical use of the technology. reimagine work, and shift how companies deliver
value in leaps versus incremental steps. Similarly,
it requires a radical approach to transforming the
operating model to unlock its full potential.
PIONEERING AI-DRIVEN As organisations shift towards an AI-ready
OPERATING MODELS Operating Model, they must design their unique
AI-transformation path. This means prioritising
By moving to an AI-ready Operating Model, organisa- flexibility over fixed structures, focusing on mindset
tions will need to chart their own AI-transformation beyond just skills, valuing data over traditional
journey by prioritising an adaptive blueprint over procedures, emphasising the evolution of human
structure, emphasising mindset more than just capital value rather than mere reskilling, and prior-
skillset, valuing data above procedure, placing itising ethical considerations over conventional
emphasis on the transition of human capital value value propositions.
ABOUT THE AUTHOR
Hamilton Mann is the Group VP for Digital Marketing and Digital Transformation at Thales. He
is also a Guest Lecturer at INSEAD and a Senior Lecturer at HEC Paris and EDHEC Business
School. Additionally, he serves as a mentor at the MIT Priscilla King Gray (PKG) Center and hosts
The Hamilton Mann Conversation (www.hamiltonmannconversation.com).

CYBER TELECOMMUNICATION
SECURITY TECHNOLOGY
UNIFIED
COMMUNICATIONS
SMART
OFFICE
MANAGED
SERVICES
© IMAGINE SDN BHD 2022. ALL RIGHTS RESERVED - RC20004682 imagine.com.bn

GETTING CLOSER
TO MACHINES WITH
MINDFUL STEPS
People have always had by Robb Wilson
a strong tendency to
anthropomorphise inanimate
objects. Who doesn't know
someone who gave a name
to their first car? But
perhaps we should give
more thought to the
relationships we enter
into with AI systems.
32 THE
THEEUROPEAN
EUROPEANBUSINESS
BUSINESSREVIEW
REVIEWJANUARY
JANUARY
- FEBRUARY
- FEBRUARY
2024 2024
F
ANTHROPOMORPHISM, INNATE
rom rocks to hammers to high-impact AND POWERFUL
drills, for thousands of years, we’ve grown
accustomed to holding tools in our hands. Now, The most memorable cars I’ve owned or ridden
more often, we’re holding them in our minds. in had names and a semblance of personality
AI-enabled tools are rapidly growing more that emerged from the experience of being
powerful and precise, and part of their predic- inside their cabins. Plus, their headlights looked
tive prowess lies in the ability to communicate like eyes and their grills grinned like mouths.
through written and spoken words. This puts Imagine how the dynamics will shift when we
us on the verge of interacting with machines in can have useful conversations with our cars. Or,
much the same way we interact with humans, when the IDW we speak with in the car is the
conversationally. Very soon, we will be having same one that we can continue speaking
meaningful, ongoing, human-like relationships with while walking into the kitchen.
with machines. The conversations we have with IDWs
According to Gartner, by 2025, can take all kinds of forms, but in
generative AI will be a workforce productivity settings it’s important to
partner within 90 per cent of compa- It’s becoming much consider how human we want these
nies worldwide (Gartner, Gartner IT easier to use anthropo- interactions to seem – how much inti-
Symposium 2023 Presentation, “We macy we want to create. With their
Shape AI – AI Shapes Us”, Mary Mesaglio morphism to fool an end ability to speak and write, to listen
and Don Scheibenreif, 16-19 October user into believing that and read, large language models
2023). Gartner is calling these partners they are interacting (LLMs) are already innately anthro-
digital teammates – I call them intel- pomorphic. They pass the Turing
ligent digital workers (IDWs) – and if with a real person, test1, with humanness to spare. They
we’re going to populate our daily lives which can happen by feel real.
with scores of them, there’s significant With ChatGPT in particular, we’ve
design or by accident.
work to be done. Organisations need already seen people using conversa-
to develop clear strategies and build tional AI as an armchair therapist2.
systems with intention. Designers need to poke I’ve also heard people who have intimate knowl-
lots of holes and develop the iterative chops to edge of how LLMs operate express surprise at
quickly plug them – or, better yet, quickly divert how intelligent they seem. Their anthropomor-
flows in safer directions. Users will need to be phic nature makes them incredibly powerful
able to understand what kinds of systems they are from a design standpoint – power that is, of
interacting with throughout their days. course, double-edged. These models can be
With the race already well underway to made to delight or deceive.
connect powerful generative models to organ- It’s becoming much easier to use anthropo-
isations and end users, business leaders will morphism to fool an end user into believing
need to move quickly but intentionally. Taking that they are interacting with a real person,
the long view and thinking solutions through which can happen by design or by accident.
to all possible ends will be difficult to balance Obviously designs that are intended to fool
against incoming waves of disruption, but it will people into thinking they are human are
be necessary. These early moments of intimacy unethical (perhaps barring designs that are
with machines will define the very nature of our intended for entertainment). In the realm
relationship with these powerful new allies. of armchair therapy, these systems have led
Let’s take a look at the ways we can connect people down dark and, sadly, deadly paths
with this new class of tool and the careful steps simply by virtue of being human enough to
we can take in getting there. seem trustworthy3.
Bottom line, anthropomorphism can happen personalisation offers opportunities to go

in designed and unintended ways. As we get beyond simply delighting users. Imagine if the
our footing in the world of digital teammates voice interface I mentioned earlier can look
and begin creating a new kind of intimacy with through your email inbox to find the plumber
our tools, letting machines behave more like you used a few years ago while also ordering
machines might be erring on the side of caution. you a ride to the airport, checking you into
your flight, and sending a note to your friend in
Chicago that your flight is on time.
That’s a home speaker use case loaded with
GETTING EVEN CLOSER high-value activity. The value of a truly smart
speaker in a business setting is far greater. Well-
OpenAI recently unveiled GPTs, customisable orchestrated conversational AI can remove so
versions of ChatGPT that users can train without much tedium from our work existence that the
having to write code. According to their blog4, quality of our lives will take a dramatic upswing.
”Creating one is as easy as starting a conversa- I bring this up for two reasons.
tion, giving it instructions and extra knowledge, Obviously there’s untold business value
and picking what it can do, like searching wrapped up in this level of personalisation.
the web, making images, or analysing There’s also another layer of intimacy that bears
data.” Even more recently, Google consideration. If IDWs are radically improving
launched their Gemini model, which your life in ways that the humans in your life
can understand text, images, video, aren’t capable of, how do we compartmentalise
and audio. Gemini can also perform their existence? Are they best friends? Are they
impressive tasks relating to maths and like family? Do IDWs become gods? The safe
physics, and it understands and gener- answer seems to be that they should be powerful
ates high-quality code. yet benevolent machines
GPTs are shareable and working on the behalf of all
deployable within organisa- Make no mistake, IDWs are being of us.
tions, which is a step forward used right now by forward- Regardless, the tempta-
in their evolution. Gemini’s tion to view IDWs as human
multi-modal capabilities are thinking orgs with full-blooded and to trust them more than
also game-changing. Still, AI strategy. Companies that are we should will be great. The
I wouldn’t classify either of still fiddling with siloed chatbots design choices we make in
these tools as an IDW or digital creating IDWs – including,
teammate on their own. A truly aren’t even on the same path. and perhaps especially, how
intelligent digital worker has we anthropomorphise them
shareable skills and can interact multimodally, – will dictate the kind of intimacy that emerges
but they also have the ability to communicate between humans and machines.
across an organisation's departments, systems,
and data. Make no mistake, IDWs are being used
right now by forward-thinking orgs with full-
blooded AI strategy5. Companies that are still VALUE IN CLOSENESS
fiddling with siloed chatbots aren’t even on the
same path. In removing the tedious tasks that take up so
This level of connectivity that IDWs provide, much of our time, conversational AI might allow
whether inside an enterprise or a household, us to have deeper interactions with the humans
is fertile soil for what the GPTs of the world in our lives. Conversational AI also allows us to
are aspiring towards: personalisation. This interact with machines in a way that’s far less

distracting than glowing screens. If we can peri- I recently had a conversation with Blaise Agüera y
odically ask machines questions and receive Arcas6, AI researcher, author, and VP and Fellow
written or spoken responses, we can spend less with Google Research, and he reminded me just
time looking at our pocket computers, which how collective human intelligence is. He spoke of
might also allow us to deepen our connections the "islanding" effect, where small populations are
with other humans. cut off from the larger societies, and their level of
At the beginning of this innovation plummets. “It
piece, I mentioned that we kind of shows you that people
The design choices we make in
frequently hold out tools are a little bit like neurons in
in our minds. You could creating IDWs – including, and a bigger brain,” Blaise said.
also say that our most perhaps especially, how we anthro- This realisation informs his
cutting-edge tools now take on AI, which he called a
pomorphise them – will dictate
hold our minds in them. bit unorthodox.
The GPT models were the kind of intimacy that emerges “I think about [AI] more
trained on nearly all of the between humans and machines. ecologically,” he said. “Their
internet. That represents intelligence is our intelli-
thousands of years of accrued human knowl- gence. The way we’ve arrived at AI is literally by
edge. So, in an abstract sense, if you’re asking training it on corpuses of human interaction. Their
ChatGPT a question on your phone, the tool is interaction with us is very, very human-like. I
in your hand and in your mind. At the same would argue it is human … It may be different from
time, your mind – or, more accurately, our us in its implementation, but it’s not different from
collective mind – is in the tool. us in its culture or its intelligence.”
www.europeanbusinessreview.com
As Blaise sees it, in the way that there’s collec- of intimacy between conversational AI and users.
tively more intelligence in a city than on an The work won’t be easy, but it’s hard to think of a
isolated island, AI can radically boost our more critical time to get something right.
capabilities. I agree, and
can envision technology
While this might sound lofty or REFERENCES
taking us to dizzying new
heights – actually bringing abstract in terms of things busi- 1. Turing test. Wikipedia. https://en.wiki-
pedia.org/wiki/Turing_test
us closer together as people ness can do right now, there are
2. ChatGPT is giving therapy. A mental
and enhancing our collec-
immediate choices that organ- health revolution may be next. 27
April 2023. Aljazeera. https://www.
tive intelligence to the aljazeera.com/economy/2023/4/27/
point where even the most isations are making right now could-your-next-therapist-be-ai-tech-
raises-hopes-concerns
daunting problems we face that are setting trajectories in a 3. Man ends his life after an AI chatbot
(like climate change, wide- 'encouraged' him to sacrifice himself
spread corruption, and

moment of massive acceleration. to stop climate change. 31 March
2023. Euronews.net. https://www.
euronews.com/next/2023/03/31/
inequality) can be solved. man-ends-his-life-after-an-ai-chatbot-encouraged-him-to-
sacrifice-himself-to-stop-climate-
While this might sound lofty or abstract in
4. Introducing GPTs. ChatGPT. https://openai.com/blog/
terms of things business can do right now, there introducing-gpts
are immediate choices that organisations are 5. Case Study: Global Fortune 50 Company achieves
making right now that are setting trajectories in a 83% CSAT Score By Automating Employee Experience.
OneReach.AI. https://onereach.ai/portfolio/case-study-
moment of massive acceleration. There are path- global-fortune-50-company-achieves-83-csat-score/
ways that business leaders can begin to follow 6. S2E23 Identity and Collective Intelligence with Blaise
Agüera y Arcas, VP at Google Research. December 2023.
that will foster a measured and responsible form YouTube. https://www.youtube.com/watch?v=xZ2EQgINEh4
ABOUT THE AUTHOR
Robb Wilson is the co-founder and CEO of OneReach.ai and the GSX creator / builder platform, the
only platform in the space that’s named as a leader by all of the most respected analyst firms, Gartner,
Forrester, IDC, Everest, etc. He co-authored Age of Invisible Machines, the first WSJ bestselling book
about conversational AI, and co-hosts the Invisible Machines Podcast. Robb has spent more than
two decades applying his deep understanding of user-centric design to unlocking hyperautomation.
In addition to launching 15 startups and collecting over 130 awards across the fields of design and
technology, he has held executive roles at several publicly traded companies. A trusted thought leader
in the realm of conversational AI and hyperautomation, Robb has played a part in creating a wide variety
of products, apps, and movies that have touched nearly every person on the planet.

THE KEY SUCCESS FACTORS

OF A POWERFUL AI FACTORY
by Jacques Bughin
More and more companies are leveraging significant paradigm shift, where companies are leveraging
data and deep machine learning algorithms, data and deep machine learning algorithms at their core.
leading to the emergence of the “AI factory” Many companies are moving towards this new AI factory
model. As with any transformation however, model, including incumbents. Companies such as Lockheed
knowing the keys to success makes the Martin3, BBVA4, Cleary Gottlieb5 and EQT Ventures6 present a
difference. This article explores those factors. variety of applications, from transactional legal processes to
the reorganisation of venture capital deal flows. The current
evolution of AI is evident, particularly in the high-tech and
1 THE AI FACTORY PARADIGM SHIFT

B2C sectors.
There are also reports of significant gains for companies
leveraging AI. UK-based Ocado Retail customers are using
The beginnings of artificial intelligence (AI) can be Ocado’s flagship deep learning model of inventory replen-
traced back to Alan Turing's visionary ideas in the ishment7 to boost product availability from 90% to over 98%.
1950s. Today, AI drives many successful businesses Webhelp, a French company founded in 2000 at the dawn of
such as Netflix’s video recommendations1, Airbnb’s the internet age as an online customer support interface,
assorted rentals, Google search, and Hubspot’s has transformed its model into an AI factory. Using an archi-
copilot software codes2. tecture based on UiPath Orchestrator, the data analytics
Although often used interchangeably, AI and factory constantly provides new, factual lead generation
digital technologies differ, with AI representing a more data, enriches information data from the corporate universe
powerful subset focused on tasks requiring human while workers can directly and seamlessly obtain and share
intelligence. With these unique characteristics, the information through a set of powerful user interfaces, and
emergence of the “AI factory” model has marked a provide user feedback to continuously improve the platform.

Webhelp reports productivity increases of up to 40%8 in some Personalisation is in turn continuously triggered
lead generation projects that were originally conducted blind, by open data collection and additive microsegmen-
outside its AI factory model. Moderna used AI algorithms9 tation, and leveraged by the move to an agile AI
to cut the development of its COVID-19 vaccine by a factor factory model.
of 20 in just 65 days, a process that would previously have
taken years. Chinese company ByteDance10 has adopted the 2. Data is the new oil, but it needs a refinery
AI factory approach with its flagship TikTok product, which The supply chain of your data is essential in the AI
automatically delivers a direct stream of short, personalised factory — you need to understand where it comes
videos to users, instead of relying on recommendations. This from, how it's stored, its traceability, and how it's
user switch makes the platform the market leader in short used. In general, dataOps is a crucial methodology
videos, despite YouTube’s leadership for at least a decade. for managing the data supply chain, but it's lacking
But as with the first era of digital transformation, trans- in the majority of companies.
forming AI doesn’t seem so easy11. What are the key elements
for success in the right AI factory? 3. People have a professional future
One of the paradoxes of AI has been the emphasis on
a work-less future14, in which AI will replace much
2 FIVE KEY SUCCESS

FACTORS
human activity. Contrary to
Digital transformation has this prevailing thinking, the
success of the AI factory still
paid off handsomely when depends on many human
1. The art of strategic renewal incumbent companies skills. As an example, BBVA's
Digital technologies are “strategic” in have linked their AI Factory Center15 counts
nature . When it comes to strategic
12
150 people among its 1,000-
change, digital transformation has transformation to a major strong analytics community
paid off handsomely when incumbent renewal of their strategy. and incorporates multidis-
companies have linked their transfor- ciplinary profiles, including
mation to a major renewal of their strategy. The same should data scientists, software engineers and developers,
be true of AI transformation. A clear case study in the media data architects, as well as “business translators” who
can illustrate this point: when the New York Times success- define the technical and commercial role of AI.
fully migrated to a “reader-first paradigm”13, with its online In fact, a comprehensive analysis of people skills
paywall model. As part of this strategic shift, the paywall and AI that we recently conducted with Accenture
is both a billing mechanism and an elaborate readership Research16 concluded that successful AI adop-
personalisation system that attracts and retains readers over tion requires the skills of the entire workforce to
the long term with the compelling content they seek out. be adjusted. We see that adapting the skills of the
entire workforce to the new AI operating model, and
training leaders to master this new operating model,
are two additional crucial elements in leveraging AI.
4. Your AI must be more trustworthy than

your current brand
Whether it’s beneficial AI, responsible AI or trust-
worthy AI17, the various terminological variants
always remind us that AI needs fundamental trust
to thrive18. As AI is automating decisions in more
and more mission-critical use cases, it’s important
to add some accountability to the whole system19.
After all, most reputable companies have faced
serious problems when using AI, for

example when Microsoft’s chatbot was
on the verge of spreading hate speech,
or when Amazon’s online recruitment
tended to favour a certain gender or
race. In general, the issue of trust can REFERENCES
1. Artificial Intelligence at Netflix – Two Current Use-Cases. 10 January 2022. Emerj.
only grow with the use of Generative https://emerj.com/ai-sector-overviews/artificial-intelligence-at-netflix/
AI, and with Large Language Models 2. GitHub Copilot: Everything You Need to Know. 31 October 2023. HubSpot. https://blog.
(LLMs) being mostly opaque, and hubspot.com/website/github-copilot
3. Accelerating Artificial Intelligence (AI) at Scale. 05 May 2022. Lockheed Martin.
relying on possible data bias. https://www.lockheedmartin.com/en-us/news/features/2022/accelerating-artificial-
intelligence-ai-at-scale.html
4. How we drive innovation at AI Factory?. BBVA. https://www.bbvaaifactory.com/
5. Innovation is radical, not 5. Cleary Gottlieb Launches ClearyX, A Platform for Highly Efficient, AI and Data-Driven
incremental Legal Services. 23 June 2022. ClearyX. https://clearyx.legal/2022/06/23/cleary-gottlieb-
Innovation relies on establishing launches-clearyx-a-platform-for-highly-efficient-ai-and-data-driven-legal-services/
6. EQT’s AI platform Motherbrain pushes the boundaries of the private markets with novel
and embedding a culture of innova- algorithm for better decision making. 05 November 2021. EQT Group. https://eqtgroup.
tion, enriched by experimentation. com/news/2021/eqt-s-ai-platform-motherbrain-pushes-the-boundaries-of-the-private-
markets-with-novel-algorithm-for-better-decision-making/
Especially in the context of AI, innova-
7. What are Stockouts? (+ How to Prevent Out of Stocks in 2024). 13 June 2023. Shopify.
tion is a question of business model. The https://www.shopify.com/retail/what-causes-a-stockout
top quartile of AI companies at scale 8. HOW WEBHELP ENTERPRISE OPTIMISED LEAD GENERATION IN
THE B2B SECTOR. Artefact. https://www.artefact.com/cases/
often engage in a platform strategy and how-webhelp-enterprise-optimised-lead-generation-in-the-b2b-sector/
collaborate with ecosystem partners. In 9. Moderna leveraging its 'AI factory' to revolutionise the way diseases
are treated. 17 May 2021. ZDNET. https://www.zdnet.com/article/
2020, for example, American Express
moderna-leveraging-its-ai-factory-to-revolutionise-the-way-diseases-are-treated/
partnered with the Indian Institute of 10. Artificial Intelligence Factory, Data Risk, and VCs’ Mediation: The Case of ByteDance, an
Technology Madras to create a “Data AI-Powered Startup. 02 May 2021. MDPI. https://www.mdpi.com/1911-8074/14/5/203
11. Now It is Time for AI Transformation. 19 September 2023. The European Business
Analytics, Risk and Technology” lab. Review. https://www.europeanbusinessreview.com/now-it-is-time-for-ai-transformation/
Microsoft has opened its AI Factory to 12. 6 Digital Strategies, and Why Some Work Better than
startups in Europe in fields as varied Others. 31 July 2017. Harvard Business Review. https://hbr.
org/2017/07/6-digital-strategies-and-why-some-work-better-than-others
as healthcare, green energies and 13. Disruptive Innovations and Paradigm Shifts in Journalism as a Business: From
agri-food. Advertisers First to Readers First and Traditional Operational Models to the AI Factory.
2022. Sage Pub. https://journals.sagepub.com/doi/pdf/10.1177/21582440221094819
There are a large set of new business 14. Artificial Intelligence, Its Corporate Use and How It Will Affect the Future of Work. 27
models emerging thanks to AI, such as June 2020. SpringerLink. https://link.springer.com/chapter/10.1007/978-3-030-46143-0_14
immersive AI, global AI Marketplace, 15. BBVA AI Factory, among the world's best financial innovation labs, according to Global
Finance. 18 May 2021. BBVA. https://www.bbva.com/en/innovation/bbva-ai-factory-
Computing, or AI as service, and a among-the-worlds-best-financial-innovation-labs-according-to-global-finance/
plethora of new models such as copi- 16. The art of AI maturity. Accenture. https://www.accenture.com/us-en/insights/
artificial-intelligence/ai-maturity-and-transformation
lots attached to generative AI. In
17. NRI 2023: Translating AI Principles Into Industry Practice. 2023. Portulans Institute.
general, the companies mastering the https://portulansinstitute.org/translating-ai-principles-into-industry-practice/
AI factory are at the frontier of exper- 18. A Unified Framework of Five Principles for AI in Society. 02 July 2019. HDSR. https://hdsr.
mitpress.mit.edu/pub/l0jsh9d1/release/8
imenting with those models. Are you 19. How can we ensure trustworthy AI?. 22 August 2022. Technology Magazine. https://
ready to follow suit? technologymagazine.com/ai-and-machine-learning/how-can-we-ensure-trustworthy-ai
ABOUT THE AUTHOR
Jacques Bughin is the CEO of MachaonAdvisory and a former professor of

Management. He is retired from McKinsey as senior partner and director of the
McKinsey Global Institute. He advises Antler and Fortino Capital, two major
VC/PE firms, and serves on the board of several companies.

Commercial and
Digital Diligence
Tech enabled data-led

VROXWLRQVIRUFRQȴGHQW Strategic Consulting
Combined Data dashboard
decision making and Digital
for originations and
Transformation
competitor monitoring
Supporting investors and brands

throughout the entire M&A lifecycle
onefourzerogroup.com
WHAT IS THE
NEAR FUTURE OF
GENERATIVE AI?
by Jacques Bughin
Generative AI has already made a big splash, but incorporating voice and vision in models like GPT4.0.
just how wet we’re all going to get in the future Looking back, the original GPT could not generally
is still a matter for speculation. Nevertheless, as produce coherent text by 2018, while a few months later,
you reach for your towel, Jacques Bughin has a GPT-2 could only follow simple instructions. GPT-3 and
few solid predictions for you. now GPT-4 now can perform a wide range of language
tasks on a par with humans.
L
Other models, such as Google’s Gemini, Nvidia’s
arge language models (LLMs), exemplified by Falcon, and Claude, have also been enhancing their
Generative Pre-trained Transformer 4 (GPT-4) performance to compete with OpenAI products.
from OpenAI, launched the year 2023, ushering in a Over the recent five years, LLMs have,
new revolution in AI and machine learning. While on average, improved their accuracy
traditional machine learning models often rely on on the multitask understanding
single-source data sets, LLMs, built on transformer (MMLU) scale, reaching human
neural network architectures, undergo training on expert-level accuracy 1. This
an unprecedented scale of compute and data. This performance has been closely
results in impressive capabilities, encompassing scaling with computational
tasks that were once exclusive to humans, such as resources. Larger models, such
reasoning, abstraction, and projection. as GPT-3, showcase significantly
As we explore the potential of these models, enhanced capabilities compared
it's essential to contemplate the future. Here are to their predecessors, leveraging
five predictions. approximately 20,000 × more data,
computation and parameters.
Another reason for LLM chasing size is that
LLMs have demonstrated a massive burst in abil-
LLMs WILL BECOME (EVEN) ities around programming or arithmetic, for a
BETTER SKILLED certain threshold of models. In general, perfor-
mance improves with scale roughly gradually and
In less than a year, LLMs have delivered an impressive predictably when the basis is the knowledge or memo-
evolution, expanding from text-based generative AI to risation component, but can exhibit “breakthrough”

behaviour when it requires much more complex data may overwhelm the training process of LLM
reasoning. Those breakthroughs are only to be and degrade the model capabilities. The challenge
seen now, and the race for performance among will be thus to find the right balance between degra-
large LLM providers competing for their share, dation and learning.
from Google to Microsoft and others, suggests that Reinforcement data techniques have proven
the performance will continue to accelerate in the to be extremely successful in the past, from
months to come. Tesauro’s system by 1995 that used reinforcement
learning to learn how to play backgammon at a
very strong master’s level, or Crites and Barto's
(1996) techniques to optimally dispatch elevators
PERFORMANCE IS “MORE in a multi-storey building. These are quickly being
WITH LESS” used for refining LLMs and supporting the accu-
racy of LLM models. The customisation of GPT is
Based on estimation by DeepMind and Meta, we one trick to make people more likely to use LLMs
estimate that the performance function increases and provide reinforcement clues, even if, today, the
by a factor of 10 for a performance of the model is
model 10 times larger not necessarily better but, in
on both parameters and
The race for performance the medium term, models are
tokens. This also means among large LLM providers sure to learn extensively from
that the scale needed to competing for their share, the feedback loops of inten-
continue improvements sive and diversifiable usage.
is challenging, and that
from Google to Microsoft A promising capability
“big is beautiful” may need and others, suggests that is the use of synthetic data,
some twist going further. the performance will which has proven its capa-
With this path, the cost of bility beside LLM in many
doing LLM is computation-
continue to accelerate in circumstances already. A
ally costly, and inevitably the months to come. publicly known case is the
raises sustainability issues, Nvidia simulator application
let alone that LLMs will have soon consumed all in its industrial metaverse that successfully lever-
the public data to train and tailor their model. ages synthetic data to train robots. In general also,
How then will performance increase in a synthetic data can be used to rebalance samples
paradigm shift involving “more with less”? We when the required prediction concerns rare events
already see a bit of segmentation happening
between some models (those above 10
billion parameters, such as Llama, Alpaca
and Falcon), and the very big ones with
hundreds of billions of parameters. The 10
billion parameter models are, more often
than not, based on OpenSource, but are also
experimenting with new methods to break
away from high computing costs.
Those models are looking at new ways around
data, such as data repetition, data augmentation,
synthetic data on top of merging generative AI with
reinforcement learning. Repetitive data helps LLM
to learn the content structure and patterns more
effectively, but there is clearly a risk that duplicate
such as financial fraud or manufacturing defects.

Generating synthetic instances of such events
increases model accuracy. Watch this space, but the
“more with less” is here to stay and expand.
CRACKING xAI FOR LLM
Explainable AI, or xAI, is the capacity of AI models

to elucidate the reasoning behind their actions in
a manner comprehensible to humans. This trans- approaches, aim to make LLM models more trans-
parency is paramount for user trust, particularly in parent. Alternative solutions involve exploring
safety-critical domains like healthcare, self-driving techniques beyond pure neural-network-based
cars, or finance. AI providers align with the need models, such as merging with symbolic AI or
for understanding flaws and biases in models to employing synthetic data.
enhance accuracy.
xAI has seen substantial progress outside LLMs,
employing attribution techniques like gradient-based
methods and SHAP values. However, challenges for LLM GOES TO THE EDGE
LLMs are twofold. First, these techniques demand
considerable computational power to explain Users have often experienced latency issues with
billion-parameter LLMs. Second, the sheer volume products like ChatGPT, and concerns about the
of content absorbed by LLMs surpasses human privacy and security of data ingested in LLMs have
capacity for absorption, complicated by the high often arisen due to their reliance on cloud computing.
non-linearity of neuronal Edge AI presents a potential
networks. This is not an solution, offering connected
issue of “black box” for
There is currently no intelligence with low latency,
users only, but designers robust technique providing as well as privacy-preserving AI
themselves. Remember a clear logic of how services at the network edge.
that the model improve- A virtuous cycle emerges
ments could not be
LLMs operate, due to the where LLMs optimise last-
anticipated at the release complexity of dealing with mile network consumption,
of GPT-3; it took some time billions of connections benefiting telecom providers
for Open AI among others through improved AI-based
to realise GPT-3 break-
between artificial neurons. scheduling and resource
“
through capabilities of allocation algorithms. This
few-shot learning, and chain-of-thought reasoning. results in decreased latency and enhanced spectral
There is currently no robust technique providing efficiency.
a clear logic of how LLMs operate, due to the Monetary considerations come into play as
complexity of dealing with billions of connections well. The value of LLMs lies in their applica-
between artificial neurons. Despite claims, recent tion, not just in training sets. This suggests a shift
research by OpenAI suggests that, at the level of towards smaller, versatile LLM models for simpler
each neuron, only 2 per cent exhibit explainable applications, such as human interactions and
power above 70 per cent. conversations. Techniques like parameter-effi-
The field of xAI is flourishing out of necessity cient fine-tuning and QLoRA can fine-tune smaller
for LLM survival. Various tools, including geometry models to perform like state-of-the-art LLMs.

This trend not only secures the demand side but business users engage ChatGPT for programming
also sees consumer tech players finding value in and marketing support. But despite all the hype,
making their devices more intelligent, rather than the real industrialisation of LLM-based transfor-
relying solely on cloud providers. The intersection mation takes time.
of 5G and LLMs is expected to create a small revo- However, the true industrialisation of LLM-based
lution, with companies like Apple, Samsung, and transformations is a gradual process. Several
others actively working on it. factors contribute to this gradual progression, with
Implications of this trend include the potential the absence of established usage policies being
for LLM models to reside at the edge, such as on a prominent one. Concerns related to security,
smartphones, enabling real-time conversational hallucination, and other issues are valid, yet the
activities like image collection and chatbot discus- underlying reality is that enterprises possess a valu-
sions. In business contexts, tailored LLMs facilitate able reserve of private data. This data can fuel the
intelligent and private access to company-specific development of domain-specific models, enabling
information, supporting the vision of widespread the integration of AI as a strategic advantage within
mobile and remote work. Additionally, robots with their enterprise.
LLM capabilities could act in real-time collabora- Major tech entities like Microsoft and Google are
tion with humans. evidently targeting the enterprise markets, lever-
aging their wealth of data. Simultaneously, other
players are entering the arena with new models as a
service tailored for the enterprise space. Singapore-
LLM WILL KICK OFF THE AGE OF based WIZ.AI’s LLMs exemplify this trend, seizing
AI TRANSFORMATION the enterprise AI opportunity.
The evolving value chain is swiftly organising
The hype aside, 2024 appears poised to witness the around vertical application providers, some integrated
transition from testing to real competitive usage with LLMs. Examples include Claude’s emphasis on
of generative AI by companies. Presently, many education, FinGPT specialising in finance, Google’s
multitask language understanding.” arXiv preprint arXiv:2009.03300.

• Hoffmann, J., S. Borgeaud, A. Mensch, E. Buchatskaya, T. Cai, E.
Rutherford, D. de Las Casas, L. A. Hendricks, J. Welbl, A. Clark, T.
Hennigan, E. Noland, K. Millican, G. van den Driessche, B. Damoc, A.
Guy, S. Osindero, K. Simonyan, E. Elsen, J. W. Rae, O. Vinyals, and
L. Sifre, “Training compute-optimal large language models,” vol.
abs/2203.15556, 2022.
• Kaplan, J. S. McCandlish, T. Henighan, T. B. Brown, B. Chess, R. Child,
S. Gray, A. Radford, J. Wu, and D. Amodei, “Scaling laws for neural
language models,” CoRR, vol. abs/2001.08361, 2020.
• Kashefi, Rojina, et al. “Explainability of Vision Transformers: A
Comprehensive Review and New Perspectives.” arXiv preprint
arXiv:2311.06786 (2023).
• Letaief, Khaled B., et al. “Edge artificial intelligence for 6G: Vision,
enabling technologies, and applications.” IEEE Journal on Selected
Areas in Communications 40.1 (2021): 5-36.
• Madaan, A., Tandon, N., Gupta, P., Hallinan, S., Gao, L., Wiegreffe, S.,
focus on healthcare, Dramatron contributing to theatres
Alon, U., Dziri, N., Prabhumoye, S., Yang, Y. and Welleck, S. (2023).
and movie scripting, or Codex powering GitHub Copilot “Self-refine: Iterative refinement with self-feedback.” arXiv preprint
for software development. arXiv:2303.17651.
• Mehra, Akshit (2023). “Data Collection and Preprocessing for Large
The era of digital transformation is now a part of
Language Models,” June, Labellerr, accessed on 16 December.
history, making way for the era of AI transformation. • Srivastava, A., Rastogi, A., Rao, A., Shoeb, A. A. M., Abid, A., Fisch, A.,
The signs are that 2024 will be the year, with compa- Brown, A. R., Santoro, A., Gupta, A., Garriga-Alonso, A., et al. “Beyond
nies positioning themselves to harness the potential the imitation game: Quantifying and extrapolating the capabilities of
language models.” arXiv preprint 2206.04615, 2022.
of generative AI, recognising it as a pivotal element in • Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.-A., Lacroix,
reshaping their business landscape. T., Roziere, B., Goyal, N., Hambro, E., Azhar, F., et al. (2023), “LLaMA:
Open and efficient foundation language models”, arXiv preprint
2302.13971.
REFERENCES • Zhao, H., Chen, H., Yang, F., Liu, N., Deng, H., Cai, H., ... & Du, M. (2023).
“Explainability for large language models: A survey.” arXiv preprint
• Balestriero, Randall, Romain Cosentino, and Sarath Shekkizhar.
arXiv:2309.01029.
“Characterizing Large Language Model Geometry Solves Toxicity
• Zhao, Wayne Xin, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang,
Detection and Generation.” arXiv preprint arXiv:2312.01648
Yupeng Hou, Yingqian Min, et al. “A survey of large language models.”
(2023).
arXiv preprint arXiv:2303.18223 (2023).
• Bughin, Jacques (2023, October 1). “To ChatGPT or not
• Zhou, D., Scharli, N., Hou, L., Wei, J., Scales, N., Wang, X., Schuurmans,
to ChatGPT: A note to marketing executives”. In Applied
D., Cui, C., Bousquet, O., Le, Q. V., and Chi, E. H. (2023) “Least-to-most
Marketing Analytics: The Peer-Reviewed Journal, Volume 9,
prompting enables complex reasoning in large language models.” In:
Issue 2.
The Eleventh International Conference on Learning Representations,
• Buchatskaya, E., Cai, T., Rutherford, E., de las Casas, D.,
2023. [https://openreview.net/forum?id=WZH7099tgfM]
Hendricks, L. A., Welbl, J., et al. (2022), “An empirical analysis
of compute-optimal large language model training”. In Oh, A.
H., Agarwal, A., Belgrave, D., and Cho, K. (eds.), Advances in 1. Hendrickx and colleagues (2021) creates the test based on 57 multiple-
Neural Information Processing Systems, 2022. choice questions spanning topics in the humanities to mathematical
• Chenghao Huang, Siyang Li, Ruohong Liu, Hao Wang, and Yize sciences, and collected from a wide array of practice questions for
Che (2023), “Large Foundation Models for Power Systems.” tests such as the Graduate Record Examination, the United States
ArXiv, December 2023. Medical Licensing Examination or the Examination for Professional
• Dettmers, A. Pagnoni, A. Holtzman, and L. Zettlemoyer, Practice in Psychology. According to the authors, human-level accuracy
“QLoRA: Efficient Finetuning of Quantized LLMs,” arXiv preprint on this test varies. They report that “unspecialised humans from
arXiv:2305.14314, May 2023. Amazon Mechanical Turk obtain 34.5 per cent accuracy on this test. (…)
• Hendrycks, D., Burns, C., Basart, S., Zou, A., Mazeika, M., real-world test-taker human accuracy at the 95th percentile is around
Song, D., & Steinhardt, J. (2020). “Measuring massive 87 per cent for US Medical Licensing Examinations”.
ABOUT THE AUTHOR
Jacques Bughin is the CEO of MachaonAdvisory and a former professor of

Management. He is retired from McKinsey as senior partner and director of the
McKinsey Global Institute. He advises Antler and Fortino Capital, two major
VC/PE firms, and serves on the board of several companies.

AI AND ETHICS
WHY THE
ETHICAL
USE OF AI
MATTERS
FOR YOUR
CAREER
by Jack McGuire, David De Cremer, Leander De Schutter, and Yorck Hesselbarth
I n the contemporary digital era, innova-

tions such as artificial intelligence (AI) are
profoundly transforming the business land-
anowski, Przegalinska, Magnuski & Gloor, 2019).
Balancing this desire for authenticity with the
allure of operational efficiency poses a chal-
scape (De Cremer, 2020). The buzz surrounding lenge, making it tempting for businesses to de-
ChatGPT, coupled with recent assertions about ceive customers by blurring the lines between
the sentience of Google's LaMDA, a large lan- human and machine.
guage model, underscore Specifically, organ-
the prominence of chat- isations nowadays are
bot technology in these Balancing this desire for authen- confronted with a reality
advancements (Adam- ticity with the allure of operational where chatbots demonstrate
opoulou & Moussiades, remarkable human-like qual-
2020; Ryu & Lee, 2018; efficiency poses a challenge, making ities (Collins & Ghahramani,
Tiku, 2022). Customer-ori- it tempting for businesses to 2021; Leviathan & Matias,
ented chatbots, an emer- deceive customers by blurring the 2018). This reality makes
gent application of this the choice to cut costs by
tech, offer unparalleled lines between human and machine. adopting human-like chat-
efficiency and cost-effec- bots a rational one. However,
tiveness, operating ceaselessly and respond- this choice is not so straightforward for organ-
ing to client inquiries in real time (Salesforce, isations to make. After all, customers prefer
Research, 2019). Yet, amidst these advantages the real thing (i.e., interactions with a human)
lies an ethical conundrum. Customers cherish over the artificial one, and therefore making the
genuine human interaction and can become rational choice requires organisations to adopt
quickly disillusioned when they realise they're a strategy of deceiving their customers by not
communicating with a bot, not a person (Ciech- disclosing to them that chatbots are used.

However, what are the risks when firms use chat- THE RIPPLE EFFECT ON CAREERS
bots without disclosure? What happens to the
reputation of organisations engaging in these First of all, our research unsurprisingly finds
deceptive acts when customers find out what is that organisations employing undisclosed chat-
really happening? And, even more important, bots are perceived as less ethical by customers
what happens to the employees working for when found out. Obviously, if you work for an
those organisations? When deception is found organisation that is seen as unethical in its use
out, organisations are likely to suffer reputa- of emerging technologies, it will affect your
tional damage, but will it work identity. If this is the
also tarnish the careers of Organisations that deceive their case, how will it affect the
their employees? Several customers by pretending to have judgements and subsequent
high-profile tech companies actions of these employees?
have faced backlash over the humans handle customer enquiries The Uber scandal involving
unethical use of emerging are judged to be unethical by both the suppression of sexual
technologies. customers and the employees harassment allegations
Consider the fallout from presents some useful
the Theranos fraud and working for those organisations. insights regarding how to
misconduct scandal. While respond to that question.
the company suffered legal Employees at Uber, even
and reputational damage, employees faced a those uninvolved, experienced that the compa-
backlash, too. Several of them reported difficul- ny's ethical breaches overshadowed their
ties in job transitions, with potential employers individual reputations and motivated many of
associating them with the scandal (Lapowsky, them to resign (Kosoff, 2017).
2021). As companies carry responsibility for To validate this idea, we ran a series of exper-
their employees, it is imperative from an imental studies where employees in a simulated
accountability point of view that they are aware company were asked to facilitate deceptive
of any potential effects on the careers of their chatbot use. Putting employees in this situa-
employees before succumbing to the allure of tion made them more likely to perceive their
deploying chatbots under a veil of deception. To organisation as cultivating a culture of making
test whether employees indeed suffer in their unethical requests to their workforce. In turn,
career prospects when the organisation they because of these perceptions, we found that
work for engages in deceptive chatbot prac- those employees wanted to quit their job more.
tices, we conducted several experimental and So, organisations that deceive their
field studies (McGuire, De Cremer, De Schutter, customers by pretending to have humans handle
Hesselbarth, Mai & Van Hel, 2023). customer enquiries are judged to be unethical
by both customers and the employees working
for those organisations. As a result, customers
will show no loyalty to those organisations, and
employees want to leave them. But where can
those employees go? Are they contaminated for
the job market? With today's rapid transmission
of information online, a company's unethical
practices can become widely known, and thus
impact employees' professional trajectories.
To study this phenomenon, we conducted
two more studies, where we assessed how those
employees are seen by recruiters. Our results
AI AND ETHICS
showed that employees that had worked for an offer two actionable takeaways:
organisation known to use chatbots deceptively The role of leaders. Leaders must
were perceived by recruiters to be less trust- recognise the lasting harm of deceptive practices.
worthy, were less likely to be offered a job, and Ethical technology use can bolster company
were given a lower salary when offered one. The reputation, morale, and customer trust.
deceptive use of chatbots therefore has wide-
spread repercussions. It harms not only the The role of employees. Employees
company, but also the people who work there. should be proactive, voice concerns
about unethical technology use, and
leave companies using deceptive prac-
tices before those deceptions are revealed.
THE RESPONSIBILITY OF TECH Communicating these concerns anonymously,
PROFESSIONALS: A CALL TO in private with your manager, or publicly in
ACTION team meetings and town hall sessions are all
useful and should be considered.
The case is clear. Tech professionals must
champion ethical AI use. The broader societal In conclusion, as AI's role in business grows,
implications of our creations cannot be ignored. its ethical use is critical. It's not merely about
Advocating for transpar- company profits; it's about
ency and ethical guidelines the careers and reputations
protects both the company's Advocating for transparency and of those who make up the
reputation and your own ethical guidelines protects both organisation. Prioritising
professional standing. The ethical AI practices isn't
findings from our research the company's reputation and just a business imperative;
your own professional standing. it's a career necessity.

REFERENCES ABOUT THE AUTHORS
Adamopoulou, E. & Moussiades, L. Jack McGuire is Jack McGuire is a Postdoctoral Research

(2020, June). "An overview of chatbot
technology". In IFIP International
Associate at the D'Amore-McKim School of Business at
Conference on Artificial Intelligence Northeastern University (Boston). He received his PhD in
Applications and Innovations (pp. Management & Organization from the National University
373-83). Springer, Cham.
of Singapore Business School and his MSc from University
Bogost, I. (2022). "Google’s ‘Sentient’
Chatbot Is Our Self-Deceiving Future".
College London. Prior to this, he was an Experimental
The Atlantic. Retrieved from: https:// Lab Manager and Research Assistant at the University of Cambridge,
www.theatlantic.com/technology/ Judge Business School. Jack's research examines the psychological
archive/2022/06/google-engi-
neer-sentient-ai-chatbot/661273/ consequences of artificial intelligence and its increasing application in
the workplace. This work has been published in Journal of Business
Collins, E. & Ghahramani, Z. (2021,
May 18). "LaMDA: our breakthrough Ethics, Computers in Human Behavior, International Journal of Human
conversation technology". Google – Computer Interaction, and Harvard Business Review, among others.
Blog. Retrieved from: https://blog.
google/technology/ai/lamda/
De Cremer, D. (2020). Leadership

by Algorithm: Who leads and who David De Cremer is currently the Dunton Family Dean
follows in the AI era. Harriman House. of D’Amore-McKim School of Business and professor of
Kosoff, M. (2017, March 20). "Uber’s management and technology at Northeastern University
President Resigns as Employees (Boston), and an honorary fellow at Cambridge Judge
Head for the Exits". Vanity Fair.
Retrieved from: https://www. Business School and St. Edmunds College, Cambridge
vanityfair.com/news/2017/03/ University. Before moving to Boston, he was a Provost
ubers-president-resigns-as-employ-
chair and professor in management at National University of
ees-head-for-the-exits
Singapore and the KPMG endowed professor in management
Lapowsky, I. (2021, August 31). "What
became of Theranos employees?".
studies at Cambridge University. He is the founder and director of
Protocol. Retrieved from: https:// the Center on AI Technology for Humankind (AiTH) in Singapore,
www.protocol.com/newsletters/ which was hailed by The Higher Education Times as an example of
sourcecode/theranos-on-trial
interdisciplinary approaches to AI challenges in society. He is one
Leviathan, Y. & Matias, Y. (2018, May
8). "Google Duplex: an AI system
of the most prolific behavioral scientists of his generation, and a
for accomplishing real-world tasks recognized global thought leader by Thinkers50. He is a best-selling
over the phone". Retrieved from: author, including “Leadership by algorithm: Who leads and who
https://ai.googleblog.com/2018/05/
duplex-ai-system-for-natural-conver- follows in the AI era?”, and his newest book “The AI-savvy leader:
sation.html 9 ways to take back control and make AI work”, which will be
McGuire, J., De Cremer, D., De published by Harvard Business Review Press in 2024.
Schutter, L., Y. Hesselbarth, Mai, K.E. &
Van Hiel, A. (2023). "The reputational
and ethical consequences of decep-
tive chatbot use". Scientific Reports,
13, 16246. Leander De Schutter is assistant professor at the Vrije
Ryu, H. S. & Lee, J. N. (2018). Universiteit Amsterdam, the Netherlands. He is interested
"Understanding the role of technology
in service innovation: Comparison
in leadership and decision-making in the workplace.
of three theoretical perspectives".
Information & Management, 55(3),
294-307.
Yorck Hesselbarth is building foundational models with
Tiku, N. (2022). "The Google
European values at Nyonic AI, contributing to digital sovereignty
engineer who thinks the compa-
ny’s AI has come to life". The on the continent. Previously, he conducted research in the field
Washington Post. Retrieved from: of human-computer interaction and led several cutting-edge AI
https://www.washingtonpost.
com/technology/2022/06/11/ projects for the German Armed Forces.
google-ai-lamda-blake-lemoine/
BUSINESS INNOVATION
Edge is an essential component of the cloud computing model

and has the potential to help enterprises increase speed of
action, reduce network costs and boost efficiency. But rapid
adoption or higher investment is not enough to get the best
value from edge computing. Integration of edge with the
enterprise’s overall cloud adoption strategy is the way forward.
EDGE COMPUTING IN EUROPE: A KEY

DRIVER OF BUSINESS INNOVATION
by Ram Ramalingam,
Teresa Tung, Nitu Kaushal
& Shalabh Kumar Singh
E dge computing, in step with an overall cloud strategy, can play a
major role in bending enterprises’ innovation curve. Much of in-
novation today comes from companies that adopt solutions informed by
the staggering amounts of data generated in branch offices, on individ-
uals’ health trackers, retail stores, remote oil rigs, manufacturing plant
KEY TAKEAWAYS sites, hospitals and even satellites. In most cases, it is inefficient to move
all of this data back to a central data centre for real-time analysis. And
83% of our survey respondents
believe that edge computing real-time complex analytics right where the data is produced – be it on
will be essential to remaining the factory floor, on an individual’s health tracker or at the store check-
competitive in the future but only out counter – have the potential to drive the next wave of performance
65% are using edge today. improvement across industries.
Super Integrators — edge
adopters that tie edge to FIGURE 1 Edge drives new value at the data source
business in transformation adop-
tion — comprise just 6% of edge
adopters across the globe and Transformation at the edge
5.5% of those in Europe, but gain Enable new edge use-cases, experiences
the most in terms of efficiency, Immersive experiences, AR/VR, and Generative AI
cost reduction and capitalising with instant response
on revenue opportunities. On-premise & cloud Secure, personalized content delivered in real-time
Centralized compute on mobile or remote devices
Our three-step framework Data sovereignty and regulatory compliance ensured
for optimal edge adoption Localization of algorithms & data to amplify human
potential, regardless of heritage
includes strategising for edge,
Smart supply chains with next-level efficiencies
scaling edge across the enter-
prise and strengthening allied
capabilities such as operations Keep workloads on-premise
or in the cloud that demand Move existing applications
and human resources. full compute power to edge that benefit from Create innovative
faster response applications at the edge
to drive new value

WHAT IS EDGE COMPUTING?
We define edge computing as a new capa-
bility that moves computing to the edge of
the network where it is closest to users and
devices — and most critically, as close as possible
to data sources. Edge computing is designed to take
advantage of the growing number of smart devices
operating at the network perimeter by enabling more
complex data processing at the point of collection.
Edge is an essential component of the cloud
computing model. When computing and analysis
are pushed away from the centralised cloud closer
to the device where data is generated, it is possible
to trigger immediate action at the point of data
For our research report, “Leading with Edge
generation. Edge increases the speed of action,
Computing,4” we surveyed 2,100 C-level execu-
reduces network costs and gives a boost to effi-
ciency. Even though Edge comes with its own set of tives across 18 countries to explore the extent of
unique security challenges, the reduced potential for edge adoption and factors that influence success.
cyberattacks during the transfer of data to the cloud We found that while enterprises are convinced
from the increased number of devices involved today about the possibilities edge enables, adoption is
means enhanced security in other areas. yet to reach optimal levels — 83% of our survey
respondents believe that edge computing will be
essential to remaining competitive in the future
We see several examples of edge adoption but only 65% are using edge today.
in many industries. Take the case of a UK-based However, as we have seen, enterprises across
international energy services company that industries are rapidly adopting edge computing
designs, builds, operates, and maintains oil, to improve performance and reduce latency,
gas and renewable energy assets. It wanted to essential for deploying levers of reinvention
increase efficiency, productivity and safety at such as AI, including generative AI. While the
its project sites during the critical construc- trend is recent, it is catching up, as our research
tion and commissioning phases. The Internet below shows, and European companies are
of Things (IoT) enabled the collection of data keeping pace with the rest of the world.
from sensors, equipment and workers. Edge
computing enabled onsite analysis of this data Edge adoption is relatively recent, with majority
to drive real-time improvements in operations, of executives beginning edge adoption between
FIGURE 2 1 and 4 years ago.
resiliency, worker safety and security.
In retail, Starbucks combines in-store IoT
capabilities with cloud computing to run real- When did your
40%
organization first 37%
time analytics on the machines used by baristas start adopting edge
36% 36%
for personalising orders and for predictive computing (e.g.,
maintenance.1 proof of concept)?
In auto and transport, Tesla has created its

17%
own chipset to custom-build a supercomputer Europe
11% 12% 12%
that trains the AI systems within the car, which Global
will rely on data from at-scale AI training across
its fleet.2 Heathrow Airport is experimenting
More than 5 years ago 3-4 years ago 1-2 years ago This year
with AI combined with 3D scanning on edge to
prevent wildlife trafficking.3 Source: Future of Edge Survey, January 2023
Total Sample; n=2100; Europe=683. Total Edge adopters: n=1372; Europe=421
(Remaining from the total sample have not implemented Edge plans yet)
BUSINESS INNOVATION
Global spending on Edge is expected to be European enterprises have taken actions to

$208 billion in 2023, a 13.1% jump from 2022. secure their sensitive data at the edge, while
Enterprise and service provider spending on 55% plan to do so in the next 2-3 years.
hardware, software and services for Edge is fore- As Europe focuses on taking manufacturing
cast to sustain this pace of growth through 2026 to the next level by establishing smart facto-
when spending will reach nearly $317 billion.5 ries, edge computing can come into play in
many ways. It can improve quality
assurance by fine-tuning auto-
Edge is one way to mation and reducing variation
EDGE IN EUROPE help companies meet in output through onsite analysis
of data. Manufacturers that need
For European companies, the data protection regula- to locate computing closer to the
enhanced data security that edge tions, such as GDPR, by site of production to enable smart,
entails is a major attraction. restricting the sharing of connected systems – which can
Concerns about data privacy, supply be cumbersome and expensive –
chain vulnerabilities and uncer- data over cloud for specific could look to edge computing for
tainty about where critical data is requirements and across a solution. Edge is also critical
stored and processed in the cloud for incorporating artificial intelli-
limited servers—all with
are driving many countries to treat gence, including generative AI, in
digital sovereignty as a regulatory, the consent of the user. industry, enabling data analysis in
responsibility and reputational real or near real-time and deliv-
matter. An increasing number of European ering key business insights.
companies are considering sovereign cloud Edge computing is also a step towards
investments to comply with numerous regula- sustainable computing. Data centres and trans-
tions, but also find it complex to implement.6 mission networks consume a significant 1-1.5%
Edge is one way to help companies meet of the world’s electricity produced today, and
data protection regulations, such as GDPR, their energy use continues to grow.7 While data
by restricting the sharing of data over cloud centres are moving towards greener practices
for specific requirements and across limited such as more efficient cooling mechanisms and
servers — all with the consent of the user. renewable energy usage, edge can contribute too
Accenture’s Sovereign Cloud Survey revealed by reducing unnecessary data traffic between
that to ensure cloud sovereignty, 40% of devices and the cloud.8 Gartner predicts that with

edge computing, just 25% of enterprise data will Integrated approach), saw better outcomes. They
need to be uploaded to the cloud by 2025.9 operated with a view that edge increases the
value of their digital core and enables the inte-
gration of artificial intelligence into their core
business. Edge helped them accelerate innova-
AIMING FOR AN INTEGRATED tion and reduced costs, improved efficiency, led
APPROACH TO EDGE ADOPTION to new revenue opportunities and enabled better
customer experience. The differentiating factor
Getting the most out of edge requires more than was the strategic approach to edge adoption and
mere installation and investment. Accenture’s its integration with a broader cloud strategy.
research shows that edge adoption challenges Unsurprisingly, the Super Integrated
conventional management wisdom. The fast approach delivered the best results. Super
adopter doesn't necessarily have the greatest Integrators comprise a small group of compa-
advantage overall. Half of the companies we nies – 6% of edge adopters across the globe, and
studied adopted edge as a standalone tech- 5.5% of those in Europe. Our research shows that
nology on ad-hoc projects (following the Ad-hoc Super Integrators are 4x more likely to achieve
or Tactical approaches in the figure below) to accelerated innovation, 9x more likely to increase
achieve quick improvements to their bottom efficiency and nearly 7x more likely to reduce
line and address immediate pain points rather costs than those that follow the Ad-hoc approach.
than strategic improvements. While this enables This is because Super Integrators made the best
learning and some smaller, faster outcomes, use of the digital core’s range of technologies – cloud,
these are not the companies that achieved the data, AI, applications and platforms – using edge to
best results. Our cluster analysis shows that make quick innovation decisions. They could also
neither early adoption led by a centralised digital enhance digital capabilities with human talent.
team nor higher investments to address specific Not all companies can or need to aspire to be
business needs led to optimal outcomes. Super Integrators, but any level of integration
The other half, which applied edge across all provides future-proof resilience as their tech
parts of their business (the Integrated or Super footprint and digital core expand.
A digital core-integrated approach to edge results in

FIGURE 3 Four approaches to adge FIGURE 4 the highest percentage of outcomes achieved
Type 1 Centralized IT-led edge deployers – High

Ad Hoc 413 respondents
5
Edge outcomes achieved
Type 2 Specific-need adopters of pre-pack- Super

4,5
Tactical aged solutions – 275 respondents
Mean of responses
Integrated
4 Tactical Integrated
Type 3 Integrates with cloud and scales widely 3,5
Integrated – 684 respondents
Ad Hoc
3
Type 4 2,5
Ties edge to business in transformative Mean of
Super responses
adoption – 88 respondents
Integrated 2
2,25 2,75 3,25 3,75 4,25 4,75 5,25
Low Digital Core Strategy High

*Of 2,100 respondents, 1,372 have adopted edge. Type 4 is a sub-set
of Type 3. **The remaining 728 respondents have yet to execute on Composite of edge/cloud integration and AI,
their edge adoption plans. We refer to this group as Pre-Adopters. data architecture, security investments
BUSINESS INNOVATION
It also launched its edge talent strategy by

HOW TO UNLOCK THE working with a leading systems integrator while
VALUE OF EDGE simultaneously establishing an edge Centre of
Excellence to build a pool of relevant internal
Edge implementations often begin with rela- talent. This approach gives the company the
tively narrow objectives but should be designed flexibility to adjust the internal/external talent
with broader applications in mind. ratio as conditions warrant.
The oil and gas enterprise mentioned earlier
believes that edge will be a key part of its broader
strategy in the future, which includes AI and
blockchain. With these technologies, it hopes to We have developed a THREE-STEP FRAMEWORK
that companies can adopt to make the best use
achieve its goal of being a net-zero energy busi-
of edge computing no matter where they are on
ness by 2050. their edge adoption journey.
It realised early on that though deploying
edge had to initially be geared towards local
1
Strategise for edge: Approach edge as a
conditions where the use case would be custom- foundational capability, not as a bolt-on. Our
ised to factors such as geographic location and research has found that edge adoption works
weather, the successful use case would need to best when it integrates well with a company’s
be scaled across multiple locations globally. broader digital strategy incorporating cloud, AI
and data.
The company has created a horizontal func-
tion that focuses on standardising use cases and
2
Scale across the enterprise: Build out edge
now has approximately 40 edge assets. It used across the enterprise on the back of cloud
multiple providers for cloud services as well as and integrate with enterprise data and AI
outside consultants for edge implementation applications with a common architecture and set
and turned to key technology original equip- of architectural principles.
ment manufacturers (OEMs) for edge-related
3
Strengthen capabilities: Ensure all
systems and devices.
employees and processes are prepared
The result: reduced unplanned downtime, for Edge. Edge-related job postings have
which can cost the company up to 10% of its jumped 62% since 2019. Human-machine collab-
revenues. oration requires an environment that nurtures
creativity, flexibility and insight. This holds true
for edge-human collaboration as well.

REFERENCES
THE FUTURE OF EDGE 1

Starbucks Turns to Technology to Brew up a More Personal
Connection with its Customers, Microsoft, https://news.microsoft.com/
source/features/digital-transformation/starbucks-turns-to-technology-
Current trends indicate that edge adoption is to-brew-up-a-more-personal-connection-with-its-customers/
2
Tesla Starts Production of Dojo Supercomputer
gaining momentum globally. By 2028, 38% of to Train Driverless Cars, The Verge, July 2023,
the edge infrastructure footprint will be in Asia https://www.theverge.com/2023/7/19/23800854/
tesla-driverless-dojo-supercomputers-production
Pacific, with 29% in Europe and 21% in North 3
HRH The Duke of Cambridge Visits Microsoft’s UK Headquarters
America.10 The full potential of technologies to Learn about Project SEEKER as Part of His Work with The Royal
Foundation, Microsoft, November 2021, https://news.microsoft.com/
like IoT and AI can only be harnessed through en-gb/2021/11/18/first-of-its-kind-multispecies-ai-model-to-detect-
the simultaneous strategic deployment of illegal-wildlife-trafficking-is-ready-to-roll-out-to-airports/#:~:tex-
t=HRH%20The%20Duke%20of%20Cambridge%20met%20Microsoft%20
edge computing. This is a major opportunity UK,work%20with%20The%20Royal%20Foundation%E2%80%99s%20
for European enterprises to be at the fore- United%20for%20Wildlife.
4
front of technology deployment and establish Leading with Edge Computing. Accenture. https://www.accenture.
com/content/dam/accenture/final/accenture-com/document-2/
themselves as lead players in their respective Accenture-Leading-With-Edge-Computing.pdf#zoom=40'
industries. 5
New IDC Spending Guide Forecasts Edge Computing Investments Will
Reach $208 Billion in 2023, IDC, February 2023, https://www.idc.com/
Most of the companies that participated in getdoc.jsp?containerId=prUS50386323
our research believe that edge will be a trans- 6
Sovereign Cloud Comes of Age in Europe, Accenture, https://www.
accenture.com/content/dam/accenture/final/accenture-com/document/
formative force — one that leads to innovative Accenture-Sovereign-Cloud-PoV-Short-2023-24-May-FINAL.pdf#zoom=40
business models over the next three years. 7
Data Centers and Data Transmission Networks,
However, the right approach is an impor- IEA, https://www.iea.org/energy-system/buildings/
data-centres-and-data-transmission-networks
tant distinguishing factor between making 8
Uniting Technology and Sustainability, Accenture, Uniting
the most of this technology versus using it Technology and Sustainability | Accenture https://www.accenture.
com/content/dam/accenture/final/a-com-migration/pdf/pdf-177/
in stop-gap solutions. Aligning edge with the accenture-tech-sustainability-uniting-sustainability-and-technology.
business strategy, integrating it with the digital pdf#zoom=40
9
core and garnering support from all stake- Computing on the Edge Can Be Transformative – But Look Before
You Leap, Forbes, March 15, 2021, https://www.forbes.com/sites/
holders is the key to successful adoption. forbestechcouncil/2021/03/15/computing-on-the-edge-can-be-trans-
formative---but-look-before-you-leap/?sh=64836186f3a5
10
LF Edge’s State of the Edge 2021 Report Predicts Global Edge
Computing Infrastructure Market to be Worth Up to $800 Billion by
The authors would like to thank Jai Bagmar, 2028, the Linux Foundation, March 2021, https://www.linuxfoundation.
Toms Bernhards Callahan and Ramani Moses org/press/press-release/lf-edges-state-of-the-edge-2021-report-pre-
dicts-global-edge-computing-infrastructure-market-to-be-worth-up-
for their contributions to this article. to-800-billion-by-2028
ABOUT THE AUTHORS
Ram Ramalingam is Teresa Tung is Nitu Kaushal is Shalabh Kumar Singh is

the senior managing managing director managing director the principal director and
director and global lead and global CTO and Europe global lead for sustainable
for software & platform of Cloud First, practice lead, technology and cloud-
engineering and intelligent data and AI, at intelligent edge, related thought leadership
edge at Accenture. Accenture. at Accenture. at Accenture.
INNOVATION
INNOVATION,
THE ART OF
ABOLISHING
DISTANCE
Building the
ATTRACT DeepTech
Ecosystem
by Hervé Legenvre
DeepTech technologies have significant

effects on the economy and are an
avenue to solve many problems
humanity faces. However, the vitality
of DeepTech projects is linked to
the innovation ecosystems in which
they are embedded. In this article,
Hervé Legenvre explores the role of
innovation in DeepTech success using
the ATTRACT programme.

I
nnovation is the art of abolishing distances. Innova- passages in this book that portray innovation as the art
tion thrives by combining distinct, diverse, and pre- of linking disparate resources and expertise are less
viously unconnected capabilities and resources to cited. He noted.
solve problems. This is particularly true for DeepTech
Innovation. ATTRACT is an initiative that created a “Many improvements have been made by […] philosophers
Pan-European DeepTech ecosystem that spans detec- or men of speculation, whose trade it is not to do anything,
tion, imaging, and computing technologies. It assem- but to observe everything; and who, upon that account,
bled around 170 DeepTech projects in an ecosystem are often capable of combining together the powers of the
comprising scientific research institutions, universi- most distant and dissimilar objects.”
ties, business networks and hundreds
of students from across Europe. AT- In today's era of high
TRACT takes the art of innovation to DeepTech solutions specialisation and continuing
its next level by steering open and require long and knowledge expansion, philoso-
interdisciplinary cooperation across phers are scientists, and men of
a diverse network of stakeholders.
uncertain develop- speculation are entrepreneurs,
ment cycles, significant but innovation remains the art
capital investment, and of combining disparate capabil-
a deep understanding of ities, resources, and expertise
INNOVATION AS THE to solve problems. This art is
ART OF COMBINING THE the underlying scientific essential to convert scientific
POWER OF DISTANT principles. discoveries into economic
OBJECTS and social benefits. Bringing
DeepTech projects to market
Adam Smith's book, The Wealth of Nations is renowned requires uniting many elements that are diverse and
for the description of the division of labour, where the distinct and initially disconnected.
production process is segmented into a series of small,
specialised tasks, each carried out by a distinct worker,
enhancing overall efficiency. Smith's illustration of this
through a pin factory is well-known. Yet, the subsequent WHAT ARE DEEPTECH
TECHNOLOGIES?
In the rapidly evolving world of technology, the term

"DeepTech" describes technologies that are rooted in
groundbreaking scientific discoveries and high-tech
innovations. These are not the apps or software we use
every day at work or in our lives; they are technologies
born out of significant scientific advancements and
discoveries. DeepTech projects can sometimes be taken
to market by new ventures but also by collaborative
projects where large companies and scientists work
together. The applications of DeepTech innovations are
often not confined to a single industry. They are often
pervasive and impact a wide array of sectors. They
have broad-ranging effects on the economy, improving
efficiency and productivity across many different
industries. They are also regarded as an avenue to solve
some of the most pressing problems humanity faces.
INNOVATION
Unlike consumer-focused apps or software, DeepTech policies. DeepTech projects can gain considerably
solutions require long and uncertain development from supportive public policies. Governments can
cycles, significant capital investment, and a deep allocate funds for DeepTech R&D. This can take the
understanding of the underlying form of grants, tax incentives,
scientific principles. DeepTech or direct investment in research
ventures often struggle with The ATTRACT initiative initiatives. Such funding helps
raising money due to their longer represents a strategic DeepTech projects overcome the
gestation periods and higher initial capital-intensive phase
research and development (R&D)
and collaborative effort of development where private
investments. Entering different to harness the innovative investment might be risk-averse
markets is a major hurdle for potential of pan-European due to the long-time horizons and
DeepTech projects. Different uncertain outcomes associated
research infrastructures
countries and sectors have their with these projects. Public
own unique culture, language, and their associated policies also support education
and business practices. communities. and training at the intersection of
technology and the economy to
enlarge the talent pool available
to DeepTech companies. DeepTech challenges and
THE ROLE OF INNOVATION opportunities transcend national borders; it is therefore
ECOSYSTEMS AND PUBLIC POLICIES key to have public policies that promote international
IN DEEPTECH SUCCESS collaboration so DeepTech projects can connect a more
global innovation network. In a nutshell, governments
With such challenges, DeepTech projects need support play a crucial role in shaping the ecosystem within
from powerful innovation ecosystems and public which DeepTech projects operate.

The vitality of DeepTech projects is linked to the
the earliest stages of technological development
innovation ecosystems in which they are embedded. to market entry and beyond and to create positive
These ecosystems bring together complementary societal impacts.
resources and expertise to nurture these projects. The ATTRACT initiative represents a strategic and
Research Institutions and universities are the seed-
collaborative effort to harness the innovative potential
beds of science and DeepTech innovation. of pan-European research infrastructures and
They are also the training grounds for their associated communities. The consor-
the next generation of scientists, tium that orchestrates ATTRACT
engineers, and entrepreneurs. activities includes Aalto University,
Large companies act as part- the European Organization for
ners for DeepTech projects, Nuclear Research (CERN),
providing the scale, the European Industrial
resources, and industry Research Management
expertise necessary Association (EIRMA),
to bring DeepTech the European Molecular
innovations forward. Biology Laboratory
Startups can also (EMBL), ESADE
emerge out of Business School, the
DeepTech projects, European Southern
they are often the Observatory (ESO), the
vehicles that take European Synchrotron
nascent technologies Radiation Facility (ESRF),
out of the lab into the the European X-Ray Free
market. Venture capital Electron Laser Facility
firms, angel investors, (European XFEL), and the
and other financing enti- Institut Laue-Langevin (ILL).
ties can also play a role when The core objective of
the time is right. These investors ATTRACT is to bridge the gap
also bring their business acumen, between fundamental research and
mentorship, and networks that can help marketable solutions by engaging enti-
DeepTech entrepreneurs. ties capable of translating high-level research
into societal and commercial gains. ATTRACT has
created a rich ecosystem that meshes Pan-European
research infrastructure, universi-
THE ORIGIN OF ATTRACT ties and industry representatives.
ATTRACT focusses Research infrastructure initia-
ATTRACT is a pioneering initiative resources on the most tives and universities are hotbeds
bringing together Europe’s fundamental for advanced scientific inquiry
research and industrial communities to
promising ideas that and technological development.
lead the next generation of detection have the potential They are equipped with state-of-
and imaging technologies. Funded by to address societal the-art facilities and are staffed by
the European Union’s Horizon 2020 researchers working at the fron-
challenges and fill
programme, it aims to help revamp tiers of knowledge across various
Europe’s economy and improve people’s gaps in the market. disciplines. Industry players bring
lives by creating products, services, a commercial perspective to the
companies, and jobs. By funding this project, the table. They can help create connections between the
European Union aims to foster innovation from generation of knowledge and its practical application.
INNOVATION
A STRUCTURED INNOVATION PROCESS comprising Research Infrastructures (RIs), academic

FOR DEEPTECH PROJECTS institutions, Research and Technology Organisations
(RTOs), and industry partners. Each successful
ATTRACT offers a structured process to identify and proposal received a one-time payment of €100,000 to
develop breakthrough technologies in the field of initiate the first steps of the project.
detection and imaging. By selecting 170 potential The second phase of ATTRACT helps bring 18
projects, ATTRACT focusses resources on the most project closer to a market-ready stage. Here, there
promising ideas that have the potential to address is a focus on the proven and most promising break-
societal challenges and fill gaps in the market. In its through technology concepts from the previous phase
first phase, ATTRACT helped validate and test these showing strong potential for scientific, industrial and
concepts through a gradual and phased approach as societal applications.
illustrated in Figure 1.
FIGURE 1
A SPRINGBOARD FOR
DEMONSTRATOR PRODUCT INTERDISCIPLINARY EDUCATION
TIME
The ATTRACT Programme also adopted an “Open

Science to Open Innovation” approach by training
students in an interdisciplinary manner. This closes
IDEA PROTOTYPE MARKET a gap in the current education landscape which often
limits students to specialised activities without inter-
disciplinary interaction or entrepreneurial
ATTRACT training. The goal is to move beyond
the traditional, siloed approach
In the context of ATTRACT, a to research and instead imbue
demonstrator is an early version students with Design Thinking
of the technology that is used and similar methodologies.
to demonstrate its potential, Using these methodolo-
performance and viability gies, groups of students
to stakeholders including coming from different
early investors or partners. horizons envision how
A prototype is a more the DeepTech projects
advanced application of the of ATTRACT can be
technology, dedicated to a used to solve real-
specific market. Prototypes world problems. This
help test and refine the fosters an entrepre-
functionality of a product, neurial and co-creation
identify issues, and present mindset among students,
the concept to stakeholders demonstrating how these
including potential users. approaches can mesh
The first phase of the ATTRACT scientific research and social
Project also serves as an initial finan- innovation. The initiative is
cier backed by the European Union designed not only to enrich the
Horizon 2020 programme. A €20 million educational experience of students but
funding from Horizon 2020 supports the selection of also to contribute to a more innovative and evenly
170 technology concepts via an Open Call to consortia distributed logic of innovation.

WHAT IS ATTRACT’S SECRET RECIPE? EXAMPLE OF PROJECTS SUPPORTED
BY ATTRACT
Innovation is the art of abolishing boundaries
ULTRARAM
and distances. Innovation often emerges from the
ULTRARAM is a new type of computer memory
combination of different and previously unrelated that uses very little energy with the speed of
elements such as specialised knowledge, practical DRAM and the non-volatility of flash memory.
skills, advanced equipment, financial backing, and It's designed to save a lot of power in places
real-world challenges across various industries. like data centres and satellites in space. The
The ATTRACT initiative is about fostering a commu- next step for the project is to test ULTRARAM’s
nity as much as it is about advancing technology. It's not ability to hold data for a long storage time and
merely a source of funding; it's a networked environment its endurance level. This could change how all
where sharing ideas and support is standard practice. kinds of devices, from tiny sensors to big data
Central to ATTRACT is the idea that a diverse range of centres, store and use data. The ULTRARAM
contributors can enable the development of DeepTech team is led by Manus Hayne from Lancaster
innovation. This community includes research insti- University, it is now supported by the award-
tutions and universities that lay the groundwork for winning startup Quinas Technology as it won
new ideas through fundamental research and provide the title of “Most Innovative Flash Memory
the intellectual and physical resources to develop Startup” in Silicon Valley.
these ideas further. Students and business networks
contribute new perspectives and energy, while public RANDOM POWER
funding, companies and venture capitalists bring the The Random Power taps into the quantum
necessary investment and business expertise to help characteristics of semiconductors to create a
secure and continuous supply of random data
turn promising ideas into market-ready products.
bits, crucial for encryption. In its first phase, the
This collaborative approach within ATTRACT is
project team which included research institu-
designed to encourage openness and interdiscipli-
tions and companies, successfully engineered
nary cooperation, making it easier for people to learn
a compact, credit card-sized device that
from one another and work together effectively. By
generates these random bits and passed US
facilitating connections among these varied elements,
National Institute of Standards and Technology
ATTRACT helps to develop nascent ideas into mature tests, confirming its practicality for real-world
projects. ATTRACT speeds up the development and use. The project led by Massimo Caccia from
application of new technologies, broadening their the Università dell’Insubria has since grown
potential impact. This approach showcases the value of to include a new company formed from the
enabling collaboration across different disciplines and original team and has broadened its expertise
sectors to drive innovation forward. by collaborating with other projects. The goal
now is to develop a range of True Random Bit
ACKNOWLEDGEMENT
The author would like to thank the ATTRACT team for their Generators that can be used across various
support and for making this article possible. This includes industries, from large-scale infrastructure to
Pablo Garcia Tello, Markus Nordberg and John Wood. The the IoT and automotive sectors, enhancing
ATTRACT project has received funding from the European
Union's Horizon 2020 research and innovative programme security with quantum-level unpredictability.
under grant agreement No. 101004462.
ABOUT THE AUTHOR
Hervé Legenvre is a Professor and Research Director at EIPM. He manages educational programmes
for global clients. He conducts research and teaches on digitalisation, innovation, and supply chain.
Lately, Hervé has conducted extensive research on how open-source software and open hardware
are transforming industry foundations. Hervé is part of the project advisory committee and part of the
Independent Committee for the Socioeconomic Studies call in ATTRACT (www.eipm.org).
STRATEGY & MANAGEMENT
WHY WE NEED
A RESOLUTION
REVOLUTION
If you've ever thought that there must be a more positive, less
by David Liddle, damaging way of resolving workplace disputes than that offered
Founder and Chief Consultant by the traditional approaches, the resolution framework could
of The TCM Group well be what you've been looking for.
T
here are very few people who haven’t experienced report from the UK conciliation service ACAS puts
some form of conflict at work at some point in their the cost of this workplace conflict at £28.5 billion
career. Managers fall out with their direct reports every year, equating to just over £1k on average for
over performance issues, team members come to blows as a every employee.
result of differing working styles, colleagues clash with each The causes of workplace disputes are complex
other over whether work has been allocated fairly or whether and varied. However, with over 20 years' experience
some are being treated more favourably than others. of mediating in some of the most complex disputes
As recent headlines graphically illustrate, we are also imaginable, I firmly believe that the way organ-
seeing an increasing amount of alleged bullying, sexual isations typically handle conflicts, complaints,
harassment, misconduct, and discrimination of all kinds. The and concerns is contributing to this rising tide of
working world is clearly not a happy place for many people. dissent, disagreement, and dysfunctional behav-
It seems that our offices, shops, and factories are becoming iour. We need a fresh approach to resolving the
increasingly confrontational, vitriolic, and toxic. myriad of issues that inevitably arise at work, one
It’s a problem that is costing industry dearly, in both that is more appropriate for today’s fast-changing
human and financial terms. Research from the CIPD suggests modern workplace.
that nearly 4 in 10 UK employees experience some kind of A growing number of forward-thinking organ-
interpersonal conflict at work over the course of a year. A isations – Burberry, HSBC, Aviva, and the BBC to

name just a few – have recognised this and are reframing punishment as its primary outcomes. This model of
the damaging and divisive disciplinary, grievance, and justice is reductive, i.e. it encourages right/wrong,
whistleblowing policies and processes of the past. They are win/lose, and blame/shame mindsets and behav-
developing an overarching resolution framework which iours. It infantilises our workforces, meaning that,
allows them to resolve conflicts and concernsncerns in
i a more when a problem arises, instead of encouraging
compassionate, people-centred, and values-driven way. employees to sit down and sort it out through face-
Says Claire Salter, Director of Global Employee Relations to-face, adult dialogue, people are typically plunged
at Burberry: “We know that conflict is unavoidable, but straight into damaging and divisive formal and
we know that when it’s handled well, it can be a driver for often deeply damaging and unsafe processes.
positive change. In fact we strongly believe that positive and This causes a huge amount of stress and
constructive conflict can fuel our creativity – and as a creative distress to the people involved. Employees become
business, creativity is at the core of everything we do. That’s increasingly anxious and upset as the sense of
what excites us about implementing a resolution framework psychological safety diminishes. As their stress
– it’s an opportunity for us to reframe the conversations that levels rise, their performance starts to suffer, with
we’re having about conflict across the business, and it gives little option but to take time away from work or,
our leaders the tools and space they need to resolve those ultimately, leave the organisation. It’s not just the
issues early, informally and constructively.” people directly embroiled in the conflict who are
So what does a resolution framework look like, and how affected. Everyone around them is impacted, too.
does it work in practice? Colleagues take sides, a tense atmosphere takes
hold, and productivity declines, so much HR and
management time is wasted.
These formal processes rarely, if ever, result in a
FROM RETRIBUTIVE TO RESTORATIVE successful outcome. Relationships are irretrievably
damaged, good people leave their jobs, corporate
The traditional processes used by organisations to manage reputation takes a hit and, ultimately, no one wins.
conflicts and concerns are adversarial in nature. They are A resolution framework, by contrast, is restor-
based on a retributive model of justice that has blame and ative in nature. As an integrated people policy
Research from the CIPD

suggests that nearly 4 in 10
UK employees experience
some kind of interpersonal
conflict at work over the
course of a year.
(IPP), it offers organisations a variety of proac- The resolution centre is a central hub which coordinates
tive and empowering approaches for securing a
constructive and lasting resolution to all types of
1 the day-to-day operation of the framework and manages
the triage process (see below) used to identify the most appro-
workplace conflict. These might include early, priate route to resolution for any conflicts, complaints, and
informal dialogue, facilitated conversations, medi- concerns that are raised. The centre also typically coordinates
ation, team facilitation, and coaching. The ability to conflict management training for front-line managers and
access more formal processes, up to and including resolution champions (see below) and, in some cases, manages
dismissal or legal action, is retained
ained for the rare in-house communication and development of tool kits and
occasions where it is identified that this is the most resources to support employees as they embrace the new
appropriate course of action. approach. The resolution centre also has an important role to
These frameworks put people before process. play in gathering data, feedback, and intelligence around the
They shift the dial towards adult-to-adult dialogue, resolution process, so that the organisation can continually
allowing organisations to develop positive work- learn and improve its conflict-management competence.
place cultures that have empathy, inclusion, and
well-being at their heart. They engender a culture The resolution index is used to triage any requests for
of cooperation and collaboration, leading to happy,
healthy, and harmonious workplaces.
2 resolution that come through to the central hub. Some of
the criteria used in the triage processs include the seriousness o
of
As Heather Palmer, Senior HR Business Partner the issue being raised, the impact of the situation on the parties,
in Culture, Policy and Employee Relations at the previous attempts to resolve the situation and their outcomes,
BBC, explained in a recent webinar, there is also a the number or frequency of previous complaints, and the risk
strong business case for adopting this more progres- of the situation to the parties and the organisation. Each request
sive, restorative approach: “We’ve used diagnostics to is scored against the criteria, and the most appropriate route to
identify that moving to a resolution framework is not resolution is assessed. The lower the score, the more informal
just about improving the people-centred experience the remedy (such as local resolution, a facilitated conversation,
in dealing with conflict resolution, but there’s a busi- coaching, or mediation). In higher-scoring incidences, organ-
ness and economic benefit in terms of the money isations may consider undertaking a detailed fact-find, which
and time that we’re spending in formal processes.” may in turn lead to a more formal resolution meeting.
GETTING UP AND RUNNING The involvement and

collaboration of key
The involvement and collaboration of key
organisational stakeholders
organisational stakeholders – senior leaders,
managers, HR, employees, and union part- – senior leaders, managers,
ners – is key to successful development and HR, employees, and
implementation of a resolution framework.
Companies who have successfully intro-
union partners – is key to
duced this new approach have typically successful development
involved all interested parties from the and implementation of a
outset in developing overarching frame-
works that work in their specific operational
resolution framework.
circumstances and align to their organisa-
tional purpose and values.
Three core elements of the resolution frame-
work are the resolution centre, the resolution
index, and resolution champions.

Resolution champions are an internal cadre of trained volunteers who
3 support the parties through the resolution process and for at least a year after,
to ensure that no further issues arise and that the agreed outcomes or improved
relationships are retained.
HOW IT WORKS IN PRACTICE
The resolution framework encompasses the following key steps:
STEP 1: AN ISSUE ARISES

The resolution framework starts with the premise that it is the responsibility of
the local line manager to resolve any situations as early and constructively as
possible. This usually takes the form of an early, informal resolution meeting,
where the people concerned are able to discuss their disagreements or concerns
openly and honestly, in a safe and supportive space. In the vast majority of cases,
this approach generates a successful resolution

and no further action is needed. Should prob-
lems start to arise again, a resolution champion
will be on hand to help nip any issues in the bud
and get everyone back on track. The success
of an early-resolution meeting does, of course,
depend greatly on the line manager's having an
understanding of the nature of conflict and how
to resolve it. Organisations cannot assume that
their managers will naturally have the confi-
dence and competence to do this; they will need
training to support them in this important role.
STEP 2: REQUEST FOR RESOLUTION

If attempts to resolve an issue locally and at restorative approach to the traditional investigation)
source are not successful, the next step is for the or a resolution meeting (as opposed to a disciplinary
parties involved to submit a "request for reso- or grievance hearing).
lution" to the resolution centre. Trained staff Organisations that have reframed their traditional
within the centre will triage the request against approaches to managing conflict and adopted a reso-
the predetermined criteria in the resolution lution framework are benefiting in a number of ways.
index, and will make a recommendation for the They are resolving issues sooner and more effectively
most appropriate route to resolution. This step and protecting workplace relationships, with a corre-
is the restorative alternative to the more retribu- sponding impact on performance, reputation, and
tive approach of “raising a grievance”. competitive advantage.
Anthony Fitzpatrick, Head of Colleague Experience
STEP 3: EARLY RESOLUTION and Employment Policy at Aviva, sums it up well: “The
Requests that are triaged at this early stage are introduction of the resolution framework changed the
most likely to be well suited for one of a number whole concept and dynamic of managing conflicts
of early-resolution methods. This might include and complaints within the workplace. Culturally,
a facilitated conversation, where both parties the change in language and emphasis have been so
are able to share their concerns, express their important. We aren’t focusing on being aggrieved, we
needs, and find a mutually acceptable way of are focusing on resolution, and that is a fundamental
restoring the relationship and moving forward. difference. We have seen that the vast majority of
Some cases will be suited to mediation, where cases are settled either at triage or through a facili-
a trained, independent mediator will bring tated conversation. This allows everyone to focus on
people together to engage in powerful dialogue their job and increases morale and motivation, which
that will help to transform their dispute from is good for personal well-being; it’s good for the busi-
dysfunctional and destructive to constructive ness and it’s good for the customer, too.”
and functional. Other options might include
team facilitation or coaching. Each approach is
suited to a particular type of conflict and its rela-
ABOUT THE AUTHOR
tive complexity or severity.
David Liddle is the founder and chief
STEP 4: FORMAL RESOLUTION
consultant at The TCM Group and author
For higher-scoring cases, the triage process of the book Managing Conflict: A practical
will lead to a more formal resolution process. guide to resolution in the workplace,
This might be a fact-find (an alternative, more Kogan Page, 2nd edition, 2023.

TM
Ridiculously Extraordinary
Digital CX AI Services Trust and Safety Consulting
For more information: www.taskus.com | info@taskus.com (888) 4000-TASK

PSYCHOLOGY
THE
PASSIVE-AGGRESSIVE
INDIVIDUAL
by Adrian Furnham
You must have met a few, or maybe

you are working with one right
now. Who are passive-aggressive
individuals and what makes them
tick? Adrian Furnham explores the
personality disorder in this article.

H ow to cure a major (mental) illness? The
cure does not involve years of painful soul
searching, psychotherapy, or complex cognitive behav-
• Passive-dependent individuals, who were
described as clingy, helpless and constantly
indecisive;
• Passive–aggressive individuals, who were
iour therapy. The cure is cheap, effective and instant.
Now that should get the attention of the NHS executives inefficient, pouty, stubborn, prone to
procrastination and very obstructive;
as well as the medical profession! The secret? Abolish
the illness category. This is what has happened to the • Aggressive individuals, who were destructive,
diagnosis of being Passive-Aggressive. irritable and resentful.
Psychiatric diagnosis has always been difficult
and unreliable. It was said years ago that the best Sixteen years later the latter two types were merged
way to cure schizophrenia in America was to bring into PAPD. It was suggested that people with PAPD
patients to England, where snipe rather than confront, and
they might be considered mask their opposition to, and
only “eccentric” or “somewhat The history of the PAPD rebellion against, authority. They
quirky”. Schizophrenia is, or was, suggests the concept shirk responsibility and sabotage
diagnosed less frequently here originated in the American others. The list of symptoms grew
than in the USA. as the manuals were updated
Christopher Hitchens pointed military soon after the to include behaviours such as
out that the Catholic Church Second World War to apparent forgetfulness, dawdling
has abolished the concept of describe difficult, childlike and intentional inefficiency.
“limbo”, the place certain souls
will dwell until Judgement Day.
soldiers who were in PAPD was described as a perva-
effect, social dissidents. sive pattern of passive resistance to
At the stroke of a pen, possibly
demands for adequate social and
accompanied by plainchant and
occupational performance, begin-
the whiff of incense, a mythical theological location
ning by early adulthood and present in a variety of contexts,
disappears. That is what has happened to Passive-
as indicated by at least five of the following:
Aggressive Personality Disorder.
1. procrastinates, i.e., puts off things that need to

THE BACKGROUND be done so that deadlines are not met
2. becomes sulky, irritable, or argumentative when
Psychiatrists have manuals that describe, in detail, asked to do something he or she does not want
to do
the behaviours associated with a range of psychiatric
3. seems to work deliberately slowly or to do a bad
disorders. This is true of what are called Personality
job on tasks that he or she does not want to do
Disorders. And every so often the manual is updated.
4. protests, without justification, that others make
From 1952 to 1987, there was a Personality Disorder
unreasonable demands on him or her
called Passive-Aggressive Personality Disorder (PAPD).
5. avoids obligations by claiming to have
The history the PAPD suggests the concept "forgotten"
originated in the American military soon after the
6. believes that he or she is doing a much better
Second World War to describe difficult, childlike job than others think he or she is doing
soldiers who were, in effect, social dissidents.
7. resents useful suggestions from others
The disorder originally described typical concerning how he or she could be more
behaviours such as shirking duty by wilful productive
incompetence. These ideas were then applied by 8. obstructs the efforts of others by failing to do
psychiatrists to those in civilian life. his or her share of the work
In the original conception, seventy years ago, three 9. unreasonably criticises or scorns people in
related types were identified: positions of authority
PSYCHOLOGY
The Freudians thought passive-aggressive behav-

iour is an immature defence mechanism due to its
negativistic and covert nature thereby contributing to
the suppression of emotional conflicts and impaired
problem-solving capabilities.
However, by the late 1980s, PAPD was dropped
because it was thought of not as a syndrome or disorder,
but as a specific behavioural response to particular
(work) situations. That is, it was situation-specific, not
a trait, a response pattern possibly with its origins in
childhood socialisation. People were taught it, and
reinforced by these behaviours which were “normal”
and “adaptive” in certain contexts.
Later the syndrome was renamed
Negativistic, but was appendicised
rather than put in the main text. Passive-aggressive feel unfairly treated, even abused.
Many of the behavioural descriptions types are not usually They are super-sensitive to their
remained the same, such as resistance rights, fairness, and exploita-
stressed. They sulk,
to routine tasks, consistent complaints tion avoidance. They also seem
about being misunderstood, sullen procrastinate, and leisurely; they believe success is
argumentativeness, scorning of all “forget” when asked not everything. They tend not to
those in authority, envy and resent- to do things they think be above middle-management
ment of the relatively fortunate, and levels because they are not ambi-
perpetual and exaggerated complaints
are not reasonable. tious enough. For them, the game
of personal misfortune. is not worth the candle.
Passive-aggressive types are not usually stressed.
They sulk, procrastinate, and “forget” when asked
DESCRIBING THE PROBLEM to do things they think are not reasonable. They are
called Passive-Aggressive because they are rarely
This personality type is very concerned about “doing openly defiant, yet they are often angry. They snipe
their own thing”. They demand the “right to be me”. rather than confront. And they are often quietly and
They demand a right to do their thing in their way and surreptitiously furious. They can be needy yet resentful
no one has the right to deprive them of it. They believe about those moods. They are in essence oppositional,
at work, and in private relationships, nobody has the not assertive. They often have downward job mobility,
right to “own them”. They like the companionship of precisely as a result of their beliefs and behaviours.
others but need strong defences against being ill-used. They are often late for meetings; they procrastinate;
They are particularly sensitive to fairness, equity, and they work at around 80% of their capacity; and they
equality: pretty hot topics these days. are very stubborn and hard to coach. They will rarely
They also do not find the workplace of great impor- directly confront others. Their prickly sensitivity,
tance. They can be good managers and workers. But subtle un-cooperativeness, resistance, and deep
they do not work overtime, take work home or worry self-absorption make them both unpredictable and
much about it. They certainly will not do any more than unrewarding to deal with. As a result, they have trouble
their contract specifies. They do not work to please building and maintaining a business team.
their boss or feel better about themselves. They are Passive-aggressives handle stress and heavy
often heard saying “It’s not my job” and they tend to be workloads by slowing down, simply ignoring requests
suspicious of all workplace authority. If their boss asks for greater output, and finding ways to get out of work.
them to work harder, faster, or more accurately, they Because they seem overtly cooperative and agreeable,

it can take a long time to realise how unproductive Miller (2008), a clinical psychologist, named
and refractory they can be. They are self-centred, they the passive-aggressive leader a Spoiler. He argued
focus on their own agendas, and they deeply believe that they feel disadvantaged, vulnerable, and as if
in their own superior natural talent and their right life has been rigged against them, so they carefully
to leisure. They believe they have nothing to prove and surreptitiously attack others. In essence, they
to themselves, are quite indifferent to feedback from carefully mask their opposition to, and rebellion
others, and therefore become annoyed and resentful against, authority. Thus, they shirk responsibilities,
when criticised or asked for extra effort. while demanding others do so too. He notes that
Colleagues need to be aware that Passive-Aggressive their ‘‘martyred mewling’’ is mixed with great skill at
types are not nearly as cooperative as they seem, and deflecting blame. They easily destroy team morale and
that they are only pretending to agree with you about generate a lot of animosity among co-workers. They
work and performance issues. Also, managers should are also, he concluded, masters of procrastination.
get them to commit to performance goals in public, in Hogan and Hogan (2001), leading business
front of witnesses, so that a community of people can psychologists, call these people Leisurely. They
hold them accountable. Social pressure won't change argued that these types ‘‘March to the sound of their
their views of the world, but it will serve to make their own drum’’; they are confident about their skills and
performance deficits less easily deniable. abilities; cynical about the talents and intentions of
others (especially superiors), and insist on working at
their own pace. They tend to get angry and slow down
Colleagues need to be even more when asked to speed up. They tend to feel
aware that Passive-Aggressive mistreated, unappreciated and put upon, and when
types are not nearly as they sense that they have been cheated, they retaliate,
but always under conditions of high deniability.
cooperative as they seem, and Oldham and Morris (2000), in their
that they are only pretending book describing all the Personality
to agree with you about work Disorders, claim the following five
traits and behaviours are clues to the
and performance issues.
presence of what they too call the
Leisurely Style. A person who reveals
a strong Leisurely tendency will,
they suggested, demonstrate more of
these behaviours more intensely than
someone with less of this style in his or
her personality profile.
They noted various characteristics:
• First, Leisurely men and women
believe in their right to enjoy
themselves on their own terms
in their own time; they value and
protect their comfort, their free
time, and their individual pursuit
of happiness.
• Second, they agree to play by the

rules; deliver what is expected
of them and no more, but expect
others to recognise and respect
that limit.
PSYCHOLOGY
• Third, they cannot be exploited and can

at all costs and rarely express disagreement?” and “Do
comfortably resist acceding to demands
that they deem unreasonable or above and you know or care what others expect of you?” They
beyond the call of duty. argue that PAPD people often have angry and confused
reports/subordinates who are deeply cynical. They
• Fourth, they are relaxed about time and
feel haste makes waste and unnecessary seem incapable of developing and sustaining alliances,
anxiety. partnerships and teams. And they have a reputation for
only giving lip service to many issues.
• Fifth, they are easygoing and optimistic that
whatever needs to get done will get done,
So, they are advised to focus on the gap between
eventually. what they say and what they feel; to try some empathy
by putting themselves in the shoes of those they
• Sixth, they are not overawed by authority.
work with; to learn how to deal with conflict; and to
Inevitably, different schools of psychotherapy study successful leaders. They concluded that PAPD
react very differently to the construct of a Passive- behaviour was “part of the leadership repertoire in the
Aggressive person, and, if they do not wish to big, slow-moving bureaucracies of the past”, but now
abolish it, how to treat it. Thus, those from a more you have transparently to align and commit. Perhaps
psychotherapeutic and psychoanalytic tradition because it was/is so common, it was thought of as
wish to explore the childhood and parental origins normal and OK.
of the disorder, while cognitive behaviour therapists
would advocate confronting and changing everyday
cognitions about how they are treated and perceived, SO, WHAT TO DO?
and thence modify their behaviours.
In their book aimed at helping CEOs understand Oldham and Morris (2000)
their “dark side”, Dotlich and Cairo (2003) noted that
They offer eight tips on dealing with the leisurely
PAPD people thrive in a culture where no one says what
passive-aggressive type:
they think. To help people self-diagnose, they ask four
questions: “Is saying one thing and doing another your 1. Accept the Leisurely person as he or she is.
standard operating procedure?”; “Do you have a private Don't approach a relationship with such a
person with the expectation of changing him
agenda you share with no-one?”; “Do you avoid conflict
or her to suit your needs. Rather, ask yourself

REFERENCES
what it is in this person that you like and
• American Psychiatric Association. (2013). Diagnostic and
are attracted to? Statistical Manual of Mental Disorders. 5th ed. Washington,
DC: APA.
2. If you are having difficulty with a Leisurely
• Dotlich D, & Cairo P. (2003). Why CEOs Fail. New York: Jossey
person, ask yourself whether the problems arise Bass.
because you two have different value systems. • Furnham, A. (2022). The Bright and Dark Side of Personality:
Perhaps you come from a push-hard, get-ahead, The Relationship between Personality Traits and Personality
Disorders. In: Lusk, D. & Hayes, T. (Eds). The Good, the Bad,
make-a success-of-yourself tradition, while the and the Human Dark Side at Work: New York: Oxford, pp
Leisurely person sees more value in doing his or 51-74
her own thing. Instead of judging one system as • Furnham, A., & Crump, J. (2017). Personality Correlates of
better than the other, ask yourself whether your Passive-Aggressiveness. Journal of Mental Health, 26:6,
496-501,
two value systems can coexist or merge.
• Hogan R, Hogan J. (1997). Hogan Development Survey Manual.
3. Be realistic. Life with a Leisurely person may Tulsa, OK: HAS.
demand more sacrifices from you than from him • Lane C. (2009). The Surprising History of Passive-Aggressive
Personality Disorder. Theory and Psychology, 19, 55–70.
or her. Can you make these without bitterness
• Miller L. (2008). From Difficult to Disturbed. New York:
or resentment? Amazon.
4. Make life easier for yourself. People with the • Oldham, J., & Morris, L. (2000). Personality Self-Portrait. New
York: Bantam.
Leisurely personality style don't automatically
• Schanz, C.G., Equit, M., Schäfer, S.K., Käfer, M., Mattheus,
tune in to what's important to you. Instead of H.K. & Michael, T. (2021) Development and Psychometric
waiting for this person to figure it out, let him or Properties of the Test of Passive Aggression. Frontiers in
her know your basic, essential expectations. Psychology, 12:579183.
• Sprock, J., & Hunsucker, L. (1998). Symptoms of Prototypic
5. Leisurely types can be stubborn about Patients with Passive-Aggressive Personality Disorder.
protecting their rights to do or to be as they Comprehensive Psychiatry, 39, 287–95.
please. Offer to assist in projects that need
doing and/or make a deal.
6. When the Leisurely person in your life starts
stalling, refusing, or forgetting, ask, "Are you
angry about something?" People with this
personality style have a hard time expressing
their anger directly.
7. Try to share in the Leisurely person's
pleasures. Observe his or her habits and
routines and join in.
8. Take good care of him or her. Leisurely
people are suckers for pampering and loving
attention.
There are many middle managers, but more likely

specialists (think IT) with this rather unattractive profile.
Their “pathology” may have served them well, even if the
burden of it has been “borne” by their long-suffering staff.
I have psychologist friends who say the disorder
should not be re-introduced to the psychiatric manuals ABOUT THE AUTHOR
because it is now so widespread, and even growing
Adrian Furnham is a Professor
as a result of its social acceptability that it is not
in the Department of Leadership
considered pathological or even unusual. Some argue and Organisational Behaviour at
modern “woke” attitudes actually endorse many PAPD the Norwegian Business School.
behaviours and beliefs. In his career he has worked with
One thing is clear: they are not ideal people to work too many Passive-Aggressive
with or for. people for his own good.
HR TRENDS
A new year is always an appropriate time

to stop and evaluate progress in any given
endeavour and, moreover, to consider any
new factors that may have entered into the
equation. In HR, as in almost every other
sector of business, such factors abound.
REVISITING
CHALLENGES
AT THE DAWN
OF 2024
by Pedro Cesar Martínez Moran to navigate under different circumstances and
and Simon L. Dolan unknown consequences. In this intricate landscape,
seeking support or assistance from teammates,
colleagues, or managers becomes essential for
effective functioning. Regardless of one's function,
THE CHALLENGE FOR INDIVIDUALS responsibilities, or role in a company, the inter-
dependence among team members is and will
Individuals are caught in what Zygmunt Bauman become crucial.
called "liquid life"1, which means that they are
deeply immersed in a consumer society, relent-
lessly pursuing satisfaction at an accelerated
pace, longing and shifting towards authenticity, THE EMERGING CONTEXT:
embracing individual uniqueness, and enhancing MANAGING HR IN A VUCA WORLD
capacity to collaborate generously with others AND OTHER ENVIRONMENTAL
(by leading, sharing knowledge, helping, or even FACTORS
teaching others) as well as with AI-equipped robots.
The demands of contemporary work life require Companies worldwide in a VUCA era face new
a significant focus on knowledge acquisition, a challenges on many fronts — economic, social,
diverse set of collaborative skills, and the ability environmental, and technological. Considering a

working career solely as an individual path should economic trends unfold, will likely be even more
be viewed as a mistake and, hence, it is essential to critical in addressing this global challenge.
recognise the interconnectedness of diverse roles
within the working environment. Individuals require Culture wars
interaction with others for their development, and The term “culture war” refers to an ongoing
continuous upskilling so that they can maintain an polarisation of society, often characterised by a left-
effective performance at work. versus-right or liberal-versus-conservative debate
largely conducted via social media channels. There
The objective of this short article is to briefly is a trend to steer audiences towards content that is
introduce some key factors that may influence the likely to confirm their biases, while also inflaming
HR arena, acting individually or collectively. feelings of injustice or inequality.
The rise of intelligent machines and their impact Turbulent times for economies
on the individual worker Hard economic times often lead governments to opt
The current momentum of accelerated digital for reducing spending on public services and utili-
transformation coupled with the growing strength ties, implementing job cuts, and causing a reduction
and ubiquity of generative artificial in living standards. Currently, there
intelligence (AI) necessitates the Individuals require is an intense campaign to manage
acquisition of new knowledge and inflation efficiently, secure the supply
competencies to maintain perfor- interaction with chain, reduce costs, and create safe
mance, efficiency, and productivity. others for their environments for doing business.
AI is becoming an integral part of development, and Inflationary pressures are exacer-
everyday life. While it’s undoubt- bating commercial strain due to high
edly driving innovation and creating continuous upskilling prices for certain products. On the
efficiencies in some fields, it’s also so that they can wages side, the decline in purchasing
causing a fair amount of fear and maintain an effective power is causing unease and
uncertainty. The threat to jobs is real, fostering conflict between employers
creating both opportunities and redun- performance at work. and workers.
dancies. The Boston Consulting Group
indicates that “the explosive popularity of ChatGPT
and other forms of generative AI suggests that we
are witnessing the start of a new order for business
and society”2. However, concerns arise regarding
whether handing over control of our lives to algo-
rithms can exacerbate divisions and inequality in
society. We need to identify the correct interaction
between people and machines.
Climate change is increasingly becoming a

political issue
The urgency of averting the catastrophic effects of
climate change is escalating rapidly. Technology is
widely acknowledged as playing a critical role, with
innovations such as clean energy and carbon capture
seen as integral parts of the solution. However,
the willingness of individuals and organisations to
take responsibility, along with the way political and
HR TRENDS
The evolution of work and employment The decision-making behaviour of older adults may
Changes in how people work will continue to affect be influenced by a tendency towards risk aversion,
many aspects of our lives and society. Although particularly when confronted with potential losses, as
some companies are implementing back-to- suggested by Albert and Duffy.6 This inclination towards
office policies, remote and hybrid working remain risk aversion could give valuable insights into the nature
at higher levels than before the pandemic. This of decision-making in ageing populations.
improves global mobility, as workers are no
longer tied to living close to employment centres. Ongoing urbanisation
However, it can also lead to increased social isola- Over the past 50 years, both social structures and patterns
tion and reduced social cohesion. In its survey on of coexistence, as well as mobility systems in large cities,
AI, McKinsey notes that “looking ahead to the next have undergone significant transformations. The European
three years, respondents predict that the adoption Commission, through its "Knowledge for Policy" website,
of AI will reshape many roles in the workforce”.3 notes that “people in search of better opportunities, such as
jobs, services, and education, have been moving from rural
The generation gap to urban areas across the world, and this accelerating trend
The wealth and property ownership gap between is likely to continue in the future. The number of people
generations will continue to drive global and living in cities has more than doubled over the last 40 years
social change. The Deloitte Global 2023 Gen Z and and is projected to reach 5 billion by 2050.” On one hand, it
Millennial Survey points out that the responses from favours productivity, while on the other hand it is “the cause
Gen Z and Millennials reflect “how the disruptive of problems associated with environmental degradation,
events of the last few years have shaped their lives public health, housing, congestion, and inequalities”.7
and views, and highlight that while they acknowledge
some positive change, they remain deeply concerned The new patterns of migration
about their futures”. The survey indicates that 49 per Economies will persist in benefiting from an influx of
cent of Gen Z and 62 per cent of millennials say that predominantly young, able-bodied, and active workers.
work is central to their identity, work-life balance” Concerns about the potential strain on utilities and public
and that “having a good work-life balance is the top services, as well as the impact of new arrivals on indige-
trait they admire in their peers, and their top consid- nous populations, will continue to fuel political divisions.
eration when choosing a new employer”.4 In advanced economies, the offer of jobs, visas, and educa-
tional opportunities will increasingly be used to address
The ageing population the skills gap. The International Organisation of Migration
World Bank forecasts reveal an ongoing progres- highlights the increasing numbers of key migration indi-
sive ageing process. According to these projections, cators, as depicted in table 2.8
people aged 65 and over will account for 10 per cent
of the world population in 20225. This represents Key facts and figures from World Migration
Table 2 Reports 2000 and 2022 respectively.
a doubling of the initial percentage, which was
around 5 per cent in 1960, as illustrated in table 1. • Estimated number of
173 million / 281 million
international migrants:
Population aged 65 and over as a • Estimated proportion of the world
population who are migrants: 2.8 per cent / 3.6 per cent
Table 1 percentage of total world population
YEAR PERCENTAGE (%) • Estimated proportion of female
international migrants: 49.4 per cent / 48.0 per cent
1960 5
1970 5 • Estimated proportion of
1980 6 international migrants 16.0 per cent / 14.6 per cent
who are children:
1990 6
2000 7 • Global international US$128 billion / US$702 billion
2010 8 remittances:
2020 9 • Number of refugees: 14 million / 26.4 million
2022 10
• Number of internally
displaced persons: 21 million / 55 million
Source: Our own interpretation based on the World Bank (2022).
Source: IOM 2022

The new definition of education
(rethinking education)
The speed of technological innovation means that oppor-
tunities are opening in industries that did not even exist
when many of today’s workers were at school. UNESCO
indicates that a new social contract for education is needed.
This means “an implicit agreement among members of
a society to cooperate for shared benefit”. UNESCO also
suggests that “the starting point is a shared vision of the
public purposes of education. This contract consists of the
foundational and organisational principles that structure
education systems and the distributed work done to build,
maintain, and refine them.”9
The new model of well-being

A schematic bidimensional model of well-being is
presented in exhibit 1.
Exhibit 1
PROSPECTIVE VIEW OF THE

FIVE HR MEGA-CHALLENGES IN
"TOMORROWLAND"
People management has transformed organisa-

tions. Their traditional transactional role is no
longer aligned with the current needs of busi-
nesses. Instead, human resources (HR) has evolved
into agents of change, requiring them to be closely
Source: Dolan S.L. The Secrets of Coaching and aligned with the business and take on more stra-
Leading by Values, Routledge, p. 2.
tegic responsibilities.
Organisations will encounter a multitude of
This model proposes expanding the concept of challenges in effectively managing the people
“successes” to encompass both organisational and indi- within their workforces in the coming years. This
vidual perspectives simultaneously, while also integrating can be categorised into the following areas:
the mental and physical dimensions of well-being. In
quadrant A, true well-being is achieved, where individ-
uals perform well while maintaining their personal,
mental, and physical health. Quadrant A represents a
win–win situation. Quadrant D is considered problem-
1 CULTURAL DIVERSITY AND THE CREATION
OF INTEGRATION POLICIES
atic, where not only does the employee perform poorly, The world offers a unique opportunity for people
but their health is also adversely affected. Quadrants B of different ages, races, religious beliefs, and
and C are also problematic, as either individual health or political ideologies to live together, without discrim-
organisational health is compromised. ination based on gender or sexual orientation. The
HR TRENDS
17 Sustainable Development Goals of As the modern driving businesses into the future with
the UN is a voluntary roadmap (THE adaptability and foresight.
17 GOALS | Sustainable Development).
workforce evolves, As the modern workforce evolves,
However, despite the interconnectedness a new HR trend a new HR trend is taking centre
and connectedness that exists in many is taking centre stage, manifested by the convergence
areas, barriers to travel and work persist of reward, culture, flexibility, and
in some countries.
stage, manifested meaningful work. No longer isolated
Organisational culture must enable by the convergence elements, these factors are working
people to be involved, generate commit- of reward, culture, together to drive employee satisfaction,
ment in the face of difficulties, and productivity, and loyalty. The modern
facilitate professional development.
flexibility, and employee is no longer driven by finan-
Additionally, in recent years, the coex- meaningful work. cial incentives alone; they seek a holistic
istence of different generations in the employment experience that aligns
same organisation has meant that each cohort has with their values and aspirations. This shift reflects
its motivations and idiosyncrasies, making it neces- a broader societal recognition that our jobs can, and
sary to establish communication, management, and should, provide more than just a pay packet.
reward mechanisms adapted to each group. As Baby
Boomers retire, Generation X (born between 1965
and 1980) is stepping into leadership roles across
many industries. Unlike their predecessors, Gen X
leaders combine traditional leadership styles with
2
THE EMPLOYMENT RELATIONSHIP
MODEL, PART-TIME WORK MANAGEMENT,
AND HYBRID PORTFOLIO MANAGEMENT
an innate understanding of technological advances, (FACE-TO-FACE AND VIRTUAL)
valuing both experience and innovation. This
unique combination has ushered in a more agile, The employment relationship model that has
tech-savvy, and holistic approach to leadership, worked for more than a century is under scrutiny.

The trend towards shorter working hours is contin- remote working, allowing individuals to create
uing. The four-day week is gaining ground. On the influential content during their working hours.
other hand, nomadic working arrangements are Hybrid working models are likely to dominate,
not left out of expectations. The return to the office impelled by a results-driven culture that rede-
has called into question the benefits of teleworking. fines performance metrics. As organisations juggle
Companies have seen a decline in engagement, between the compelling benefits of remote work
integration, and communication in their teams. with the intangible benefits of in-office collabora-
The physical distance has not been reduced by the tion, one thing is certain: the future of work will be
virtual relationship. The gig economy or on-de- a crucible for broader societal change, forcing us to
mand economy continues to make inroads with redefine not just how we work but why people work.
new ways to generate new revenue.
The impact of changes in the relational models
of work would affect the degree of participation and
mobilisation in the pursuit of a common project.
New knowledge will make it necessary to recycle
certain foundations, update others, and incorpo-
rate new skills and capacities.
Today, people do not come to the office to work
alone in a corner. They come to participate in
collaborative projects, solve operational problems
more quickly, and contribute their ideas and exper-
tise to strategic projects. Spaces and tools, both
physical and digital, for co-creation and collabora-
tion are needed.
As an example of the changing times, employees
who live within 80 kilometres of a Zoom office in
the United States and beyond will be required to be
present at least two days a week from September
2023. Zoom describes this as adopting a “structured
hybrid approach”.10
A significant number of employees, including
70 per cent of Gen Z and half of millennials, are
turning to side hustles and gig work. Their motiva-
tions are twofold: the need for extra income due to
3 DEMOGRAPHIC CHALLENGE AND A NEW
DEFINITION OF TALENT
the rising cost of living and inflation, and a prefer- The earth’s inhabitants are experiencing a contin-
ence for flexible, skills-based jobs that deviate from uous and steady growth in resource consumption.
the traditional 9-to-5 model. This growth is closely tied to the increasing concen-
Traditional side hustles, such as driving for a tration of populations in urban centres, driven
transportation company or delivering food, have by the generation and demand for employment
historically supplemented incomes. However, the opportunities in these areas. However, depopu-
modern gig economy has been transformed by the lated and ageing regions face a challenge known
explosive growth of influencer marketing. The influ- as the “talent development trap” in Europe. This
encer market is estimated to reach $21.1 billion in phenomenon highlights the need to reconsider the
2023, a 29 per cent jump from 2022. What started as definition of tomorrow´s talent and emphasises the
a trend among the youth is now engaging a broader importance of lifelong learning.
demographic with popular platforms. This shift is The maintenance of trades and professions
being fuelled by the accessibility and flexibility of is undergoing changes influenced by shifts in
HR TRENDS
consumption patterns and the evolving habits

and behaviours of both companies and individ-
uals. Individuals are calling for the creation of fair
and non-precarious employment opportunities,
emphasising the need for talent to be recognised
and promoted without bias or discrimination based
on gender, age, sexual orientation, or disability.
One of the most pressing paradoxes in the world
of work is the challenge of positively correlating
labour supply and demand. There is a growing need
to identify and attract talent to fill the new profes-
sions and roles that are emerging, addressing the
evolving dynamics and expectations in the contem-
porary workplace.
Talent is not perceived as an abundant resource
but rather as a scarce commodity. Managing
talent becomes crucial, especially when there is a
need to fill vacant positions in the face of a lower
supply compared to the potential demand. The
ongoing war for talent is expected to intensify,
particularly with the increased involvement of Al and aligns with sustainable and socially responsible
recruitment processes. practices, contributing to the organisation's overall
The competition for human capital will influence success and impact.
the speed at which individuals join and integrate into
organisations. This dynamic will compel companies
to consistently enhance their talent management
processes to attract, retain, and effectively integrate
skilled individuals into their teams.
4 EMPLOYEE WELL-BEING AND THE ACTIVE
PROMOTION OF RESILIENCE
Contemporary organisations are no longer Well-being serves as an antidote to stress, and indi-
solely based on the production of goods and viduals have become increasingly aware of the
services but also consider the importance of taking care of
economic and social impact their health. Companies that
of those decisions. This shift
Talent is not perceived invest in the well-being of their
encourages an approach rooted as an abundant resource workforces often experience an
in sustainable talent manage- but rather as a scarce increase in overall performance.
ment and development, fostering
greater stakeholder engagement
commodity. Managing talent However, stress levels have also
seen a rise, emphasising the
with a more pluralistic focus. becomes crucial, especially need for resilience, which can
Digitalisation has provided when there is a need to fill help mitigate stress and foster
organisations with tools not only a healthier workplace envi-
for recruiting the best talent but
vacant positions in the face ronment. As Casademunt and
also for managing, developing, of a lower supply compared Dolan state, “resilience can act
and retaining that talent. In this to the potential demand. as a protective factor for avoiding
context, organisations should issues of poor mental health.”
prioritise efforts to build a positive reputation as an They have developed a “resilience core rings model”.
employer brand. This entails creating a workplace This means a “process of six stages where some of the
culture that is attractive to potential employees most relevant research in the areas of neurobiology

The integration of AI requires an expansion of
employee skill sets, resulting in the emergence of
new roles and responsibilities and, consequently,
new jobs. Individuals must adapt to the digital
skills essential for survival, life, and work. The
European Union's digital competence framework
for citizens underscores the importance of not
only knowledge but also skills and attitudes in
engaging with the world.
Despite concerns about potential job loss due
to the rise of AI, the tech-savvy Gen Z is particu-
larly open to this science of making machines that
can think like humans and perceiving it as a career
accelerator. This suggests a future workforce that
embraces and integrates technology, reflecting a
positive stance on the opportunities presented by
technological advancements.
and behavioural therapy are applied to build resil-

ience and recover mental health, with no need for THE CHALLENGE OF OFFERING
any pharmacological intervention”.11 QUALITY OF WORKING LIFE AND
In recent years, HR professionals have faced ENHANCED WELL-BEING TO
mounting pressures and demands, resulting in EVERYONE
a concerning trend of HR burnout. HR burnout
is not merely an issue of individual fatigue; it The challenges mentioned above are not confined
extends to impact the overall efficiency of entire solely to people management. Rather, they stem
organisations and can pose challenges to talent from broader societal changes and their inherent
retention and acquisition. As companies expand, evolution. Technological advancements, involving
their human resource needs grow, underscoring increased investment and confidence in future
the urgent need for businesses to recognise HR progress, demand efforts in preparation and the
burnout promptly to ensure the well-being and acquisition of new knowledge.
effectiveness of their teams. Individuals need to recognise that the future
will unfold in a digital environment rather than an
analogue one. The proliferation of social relations
facilitated by information and communication
5 THE USE OF COMPLEX TECHNOLOGICAL
RESOURCES SUCH AS AI technologies should promote more extensive and
thoughtful interactions. While technology facil-
Major developments, such as the use and appli- itates business transactions such as trade, it is
cation of AI, process automation, and large-scale crucial to prevent the social isolation that can arise
data analysis, are reshaping the skills required in from excessive reliance on teleworking.
the workforce, necessitating continuous knowl- Improved living conditions have led to the
edge updates. These skills are broadly categorised convergence of different generations in the same
into cognitive, digital, self-leadership, and inter- workplace. Collaboration among these diverse
personal, reflecting the imperative to adapt to the generations is increasingly vital for organisa-
evolving challenges in the world of work. tional success.
HR TRENDS
CONCLUSION only contributes to a healthier and more engaged

workforce, but also fosters a positive work environ-
Private as well as public companies are compelled ment, ultimately leading to increased productivity
to confront the ongoing challenges posed by global and quality of life for all employees.
evolution. They cannot afford to ignore the changes
occurring around them. To thrive in this dynamic
environment, companies must adopt adaptive and NOTES AND REFERENCES
flexible organisational models. Articulating and 1. Bauman, Z. (2005), Liquid life, Polity.
implementing congruent shared values (i.e. sustain- 2. Boston Consulting Group, "AI at Work: What People Are Saying",
able culture) and nurturing trust will play a crucial https://www.bcg.com/publications/2023/what-people-are-saying-
about-ai-at-work, 7 June 2023 (accessed 14 December 2023).
role in gluing employees, customers, suppliers, and
3. McKinsey, "The state of AI in 2023: Generative AI's
members of the community together. It is essential breakout year" (accessed 14 December 2023). https://www.
to recognise that the ultimate beneficiary of these mckinsey.com/capabilities/quantumblack/our-insights/
efforts is the individual being, contributing to a the-state-of-ai-in-2023-generative-ais-breakout-year
4. "Deloitte Global 2023 Gen Z and Millennial Survey", file:///C:/
harmonious and sustainable workplace.
Users/HP%20Pavilion/Downloads/deloitte-2023-genz-millennial-
In the era of automation and digital trans- survey%20(1).pdf (accessed 18 December 2023).
formation, the dispelled significance of human 5. World Bank, https://datos.bancomundial.org/indicator/
creativity, empathy, and collaboration has become SP.POP.65UP.TO.ZS (accessed 18 December 2023).
6. Albert, S. M., & Duffy, J. (2012), "Differences in risk aversion
increasingly evident. Forward-thinking compa-
between young and older adults", Neuroscience and
nies acknowledge this and invest in cultivating a neuroeconomics, 3-9.
culture that values these human elements. This not 7. European Commission, "Knowledge for policy", "Megatrend
only contributes to the company’s growth but also continuing urbanisation", https://knowledge4policy.ec.europa.eu/
continuing-urbanisation_en#megatrendcontinuingurbanisation
nurtures a workspace where individuals feel genu-
(accessed 18 December 2023).
inely appreciated and motivated. 8. International Organisation for Migration, "World Migration Report
The concept of total well-being encompasses 2022", file:///C:/Users/HP%20Pavilion/Downloads/wmr-2022_0.pdf,
various dimensions, including physical, mental, (accessed 18 December 2023).
9. UNESCO, "Rethinking the futures together", https://courier.unesco.
emotional, social, and financial health. By actively
org/en/articles/rethinking-our-futures-together (accessed 18
supporting employees’ well-being in all these areas, December 2023).
organisations can enhance productivity, reduce 10. Zoom blog, "How to build your best hybrid work environment
absenteeism, and elevate employee morale. with Zoom", https://www.zoom.com/en/blog/best-hybrid-work-
environment/ (accessed 18 December 2023).
And finally, adopting a holistic approach to
11. Javier S. Casademunt and Simon L. Dolan, “The Resilience Rings:
actively promoting employee well-being can be A New Neuropsychological Framework for Building Resilience”,
regarded as a strategic move for organisations. It not The European Business Review. March 2023.
ABOUT THE AUTHORS
Dr Pedro Cesar Martínez Moran Dr Simon L. Dolan is a professor at Advantere

is the Director of Master in Talent School of Management and the University of
Management at Advantere School Comillas. He is the former director of the ESADE
of Management. Since 2017, he has Future of Work Chair. He has a PhD in People
also been the Director of Master Management and Work Psychology from the
of HR at the Pontificia University University of Minnesota and is a former full
of Comillas. In addition to academic work, he has professor at ESADE, the University of Montreal, McGill University,
worked in different roles as senior executive and Boston University, and others. He has published over 80 books,
senior consultant. Currently he is a member of the including academic textbooks in HR, in English, French, and
board of the Global Future of Work Foundation Spanish. He is the co-founder and president of the Global Future
(www.globalfutureofwork.com). of Work Foundation (www.globalfutureofwork.com).

Digital
collaboration
& document sharing
Board work and due diligence
Through our cloud-based solution,
users can efficientl\ access, share,
discuss and act on business-critical
information.
admincontrol.com
MAKE AI ACCESSIBLE TO EVERYONE!
Bring AI closer to business by democratizing AI
Broaden data analytics and AI/ML Accelerate AI adoption across an

avenues to anyone in the organization organization through reference
with no-code AI pre-built use cases
Improve agility, elasticity, and scalability Ensure transparency and interpretability

through HyperSense’s cloud-native, with fair usage of AI, to eliminate biases
SaaS-based approach through Explainable AI capabilities
Foster innovation and growth by

operationalizing AI across the
enterprise
HyperSense Platform is a brainchild of Subex, a pioneer in enabling Digital Trust for businesses across the
globe. Subex helps businesses embrace disruptive changes and succeed with conƬdence in the digital world.
Visit: hypersense.subex.com or write to hypersense@subex.com

The European Business Review - January-February 2024

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The European Business Review - January-February 2024

Uploaded by

Copyright:

Available Formats

The

INTRODUCING THE CONCEPT

USA $22 EU €17.5

ARTIFICIAL INTELLIGENCE AI AND ETHICS

18 The Path towards Trustworthy AI is BUSINESS INNOVATION

empowering communication globally

The concept of “artificial integrity”

4 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

THE PATH FOR

holistic AI framework. It seeks to create a unified oper-

Furthermore, while AI ethics recognises the

6 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

As AI ethics discusses accountability and

8 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

10 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

Here, human skills and decision-making are at the

12 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

Lastly, and especially in this quadrant, ethical considerations

14 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

16 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

constantly evaluated against the imperative to nurture and

Upholding artificial integrity is paramount, as

In the frenzy to champion the

18 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

20 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

THE BATTLE AGAINST AI’S INHERENT

Red-teaming, a process of “breaking” AI to understand

22 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

ABOUT THE AUTHOR

ENHANCING OPERATING MODELS'

24 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

CULTURE: MINDSET OVER

human need for stability.

The transition from traditional processes

26 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

and adaptable AI models. This is achieved by

ORGANIZATION: “BLUEPRINT OVER STRUCTURE”

CULTURE: “MINDSET OVER SKILLSET” PEOPLE: “HUMAN CAPITAL OVER RESKILLING”

28 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

can effectively cushion the impact of AI-induced

to merely auxiliary roles might prevent us from TECHNOLOGY: ETHICAL STANDS

Photo credit: Freepik.com

ABOUT THE AUTHOR

30 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

© IMAGINE SDN BHD 2022. ALL RIGHTS RESERVED - RC20004682 imagine.com.bn

Bottom line, anthropomorphism can happen personalisation offers opportunities to go

34 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

spread corruption, and

ABOUT THE AUTHOR

36 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

THE KEY SUCCESS FACTORS

1 THE AI FACTORY PARADIGM SHIFT

38 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

2 FIVE KEY SUCCESS

4. Your AI must be more trustworthy than

serious problems when using AI, for

ABOUT THE AUTHOR

Jacques Bughin is the CEO of MachaonAdvisory and a former professor of

40 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

Tech enabled data-led

Supporting investors and brands

42 THE EUROPEAN BUSINESS REVIEW JANUARY - FEBRUARY 2024

such as financial fraud or manufacturing defects.