Professional Documents
Culture Documents
Today’s Choices
Nanotechnology, Artificial Intelligence and Robotics;
A technical, political and institutional map of
emerging technologies.
www.greenpeace.org.uk
ISBN 1-903907-05-5
Published by Greenpeace Environmental Trust
Canonbury Villas, London N1 2PN
ISBN 1-903907-05-5
2.6 Discussion 41
Table 2: Applications for new materials and devices resulting from self-assembly
and self-organisation.
Table 5: Top five government spending on nanotechnology in the Far East in 2002.
Table 12: Summary of application areas for nanoscale pharmaceuticals and medicine.
Table 14: US historical funding for technology transitioning into the marketplace.
2
Abbreviations
and Acronyms
4
research and development (R&D), the main At the time of commissioning this report,
players behind these developments, and the civil society critiques of the immense R&D
associated incentives and risks. and commercial efforts taking place in
nanotechnology were quite sparse, but
New Technologies in Context already there are signs that this is changing.
Beyond the contents of this report, the In the wake of the furore over genetic
political and social processes surrounding the modification, the idea of a ‘public debate’
introduction of technologies are very about new technologies is in vogue, but this
important. For example, compare the public has to be meaningful or it will simply
response to GM crops in Europe to the wide promote cynicism.
acceptance of mobile phones. The ‘social
constitution’ of the technology appears key If public dialogue on science is to mean
to its acceptability. This social constitution anything, the approach of nanotechnology is
provides the answers to questions such as: a huge opportunity. Instead of waiting for
potential adverse reactions, the scientific
• Who is in control? community could be proactive. Why not hold
a citizens jury to determine scientific
• Where can I get information that I trust? priorities on nanotechology? From each of
the agricultural, defence, energy,
• On what terms is the technology being pharmaceutical, and information technology
introduced? (IT) sectors (and the numerous cross-overs),
the jury could examine current research and
• What risks apply, with what certainty, and its potential. It could suggest which areas
to whom? need to be highest priority. It would look at
the potential short- and long-term
• Where do the benefits fall? applications and the ‘blue skies’ element
necessary for any research programme.
• Do the risks and benefits fall to the same Research councils such as the Biotechnology
people (e.g. mobile phones are popular, and Biological Sciences Research Council
while mobile phone masts are not)? (BBSRC) and the Engineering and Physical
Sciences Research Council (EPSRC) in the
• Who takes responsibility for resulting UK could commit to considering results and
problems? utilizing the insights from the findings of
such a jury. If dialogue between science and
The evidence presented in this report suggests society is to be more than just a sophisticated
that, depending on the development pathway, means of engineering user-acceptance,
some aspects of nanotechnology might get a research councils must adopt this kind of
rocky ride, as its social constitution is more participatory initiative to allow ordinary
like that of GM crops than mobile phones. In people to have a say in the types and
particular, future disputes surrounding new trajectories of technological innovation.
technology seem certain in the light of
globalised, rapid technology transfer. The Nanotechnology
general public is also increasingly unwilling The most common definition of
to accept the word of a company or nanotechnology is that of manipulation,
Government (on the basis of brutal observation and measurement at a scale of
experience), on the risks and benefits of less than 100 nanometres (one nanometre is
technology, particularly as science and one millionth of a millimetre). However, the
commerce become more closely linked. emergence of a multi-disciplinary field called
The difficulty in making predictions about All of this suggests that the development of
the future is that R&D could still take nanotechnology will go through various
several different directions, and the materials different stages, and thus societal debate will
and processes being developed are need to be an ongoing process rather than a
technology-pushed rather than market-led. single outcome. There will need to be
After the hype about possible applications, continual incorporation of the insights from
the first real nanotechnology products are such a debate into policy and product
starting to appear in the semiconductor development as the prospects become more
6
tangible. Already some concerns are flagged up by the ETC group, and this
becoming evident. Some new materials may implies the pursuit of income streams from
constitute new classes of non-biodegradable those already possessing disposable income.
pollutant about which we have little Is the future of nanotechnology then, a
understanding. Additionally, little work has plaything of the already-rich? Will the much
been done to ascertain the possible effects of talked about ‘digital-divide’ be built upon,
nanomaterials on living systems, or the exacerbating the inequities present in current
possibility that nanoparticles could slip past society through a ‘nano-divide’?
the human immune system. Carbon Nanotechnology can only be made available
nanotubes are already found in cars and to the poor and to developing countries if the
some tennis rackets, but there is virtually no technology remains open to use. Already a
environmental or toxicological data on them. company in Toronto has applied for patents
Despite this, of the US$710 million being on the carbon molecule
spent by the US Government on Buckminsterfullerene. If ownership of
nanotechnology, only US$500,000 is being molecules is allowed, the nanotechnology
spent on environmental impact assessment, techniques for the precise manipulation of
even though a major feature of the product atoms open up a whole new terrain for
pipeline is that it consists of new materials. private ownership. As with genetic
Current proposals at EU level on synthetic engineering where genes have become
chemicals regulation are belatedly ensuring controlled by patents, things that were once
that a rule of ‘no data, no market’ will apply considered universally owned could become
to the basic information about hazardous controlled by a few.
properties of such chemicals. Knowing the
basics about the dangers of new materials is Artificial Intelligence and Robotics
a pre-requisite for effective environmental Unlike the situation for nanotechnology,
responsibility. From the Greenpeace researchers in artificial intelligence (AI) feel
perspective, this suggests that whilst ‘societal that their work has suffered because of
debate’ is highly desirable, it is a bit of a ‘public discussion’ – hype might be a better
luxury if the same old mistakes are being term – in the 1960s and 1980s which
repeated by a new generation of adversely affected advances in the field after
technologists. There is no need for grand, the delivery did not live up to expectations
new mechanisms of public involvement to and funding dropped. Many researchers now
point out the blindingly obvious. With cause feel that the goal of mimicking the human
for concern, and with the precautionary ability to solve problems and achieve goals in
principle applied, these materials should be the real world (the so-called ‘strong AI’) is
considered hazardous until shown otherwise. neither likely nor desirable because a long
series of conceptual breakthroughs is
Still other concerns are evident in the social required. Instead the focus is on ‘weak AI’ –
arena that revolve around the uses to which applications that model some, but not all,
the new technology is put – closely linked aspects of human behaviour.
with ownership and control. One possible
dystopian future would be the shift of the The number of applications for weak AI is
control of nanotechnology towards being growing. AI-related patents in the US
driven by military needs. This report does increased from 100–1700 between 1989 and
not generally support such a prospect at 1999, with a total of 3900 patents
present, although military interest in mentioning related terms. AI systems are
nanotechnology is considerable. generally embedded within larger systems –
Alternatively, corporate control has been applications can be found in video games,
8
misuse of the data are clear. Furthermore, the Equally, nanotechnology may converge much
use of computer systems for the US National sooner with biotechnology as it uses the tools
Missile Defence, and possibly for UCAVs, and structures of biological systems to
has created a different moral dilemma in that generate tiny machines. Although the
“they will be the first machines given the prospect of general assemblers may be quite
responsibility for killing human beings distant, self-replicating ‘machines’ that use
without human direction or supervision”. the tools of biology – and look more like
living things than machines – might be closer
AI and robotics are likely to continue to creep at hand through the convergence of bio- and
into our lives without us really noticing. nanotechnologies. ‘Grey goo’ might not be a
Unfortunately, many of the applications realistic prospect; ‘green goo’ may be closer
appear to be taking place amongst agencies, to the mark – quite how close is difficult to
particularly the military, that do not readily judge on the basis of the evidence in this
respond to public concern, however well report. Any creation that posed the prospect
articulated or thought through. of being self-replicating would need to be
handled with immense care to ensure
The Future environmental protection.
Nanotechnology and AI/robotics, together
with biotechnology, may well be on a Whether any of the technological futures
convergent path. In 2001 the National being scoped out in laboratories are what
Science Foundation held a large workshop to our general public would like is a question
look at the implications of this convergence that can only be answered by asking them. If
and the implications for human abilities and those concerned with the development of
productivity. AI could be boosted by new technologies, and nanotechnology in
nanotechnology innovations in computing particular, are convinced that the benefits
power. Applications of a future they hope to generate will withstand scrutiny
nanotechnology general assembler would they should have no concerns about engaging
require some AI and robotics innovations. and winning public support.
Acknowledgements
Many thanks to my supervisors, Timothy Olivier Bosch of the International Institute
Foxon and Robert Gross, Imperial College for Strategic Studies (IISS) for allowing me to
London, for their guidance and advice in interview them.
completing this report; to Douglas Parr,
Greenpeace, for commissioning the work; Finally, I am grateful to Jon Glick of the
and to Ken Green, University of Manchester American Association for Artificial
Institute of Science and Technology, for his Intelligence (AAAI); Andre Gsponer of the
review and commentary. Independent Scientific Research Institute
(ISRI); Hope Shand of the ETC Group; and
In addition, I would to thank Gareth Parry, Loretta Anania, Ramon Compano, and
Jenny Nelson, and Murray Shanahan of Jakub Wejcher of the European Community’s
Imperial College London; Abid Khan of the Future and Emerging Technologies
London Centre for Nanotechnology; and programme (EC FET) for their assistance.
10
One problem with many of the hundreds of UK Strategy for Nanotechnology provides a
documents written about emerging useful introduction to the field. In addition,
technologies every year is that they do not Ramon Compano (2001) of the European
distinguish between science and science Commission; Professors J.N. Hay and S.J.
fiction, let alone the desirable and Shaw (2000) of the University of Surrey and
undesirable in terms of ethics, choice and Defence Evaluation and Research Agency
safety (Ho, 2002b). Thus, Sections 4 and 5 (DERA); Paul Holister (2002) of CMP
aim to deal with some of these issues: Section Cientifica; Ian Miles and Duncan Jarvis
4 separates out some of the hype from the (2001) of the National Physical Laboratory
more visionary but solidly placed (NPL); and Ottilia Saxl (2000) of the
applications, whereas Section 5 provides an Institute of Nanotechnology have been used
account of the potential environmental and extensively for construction of summary
social risks that such uses could pose in the tables. Finally, the National Science
future. Finally, Section 6 highlights some of Foundation (NSF) report, Societal
the key messages of each part. Implications of Nanoscience and
Nanotechnology, supplies good information
1.3 Key references on a wide range of issues (Roco and
This report has been compiled by consulting Bainbridge, 2001). For the section on AI and
a wide variety of sources across the entire Robotics, Barbara Grosz and Randall Davis
spectrum of the debate, from industry – President and President-Elect of the
advocates to environmental and social American Association for Artificial
pressure groups. In doing so, a number of Intelligence (AAAI) – and Daniel Weld of the
sources have been particularly important. For University of Washington provide some
the section on nanotechnology, the DTI’s useful technical information.
(2002) New Dimensions for Manufacturing:
12
this underlying research into economically 2.2 Research and Development
viable products: the US NanoBusiness
Alliance, the Europe Nanobusiness 2.2.1 Introduction
Association, and the Asia-Pacific The absence of a universally accepted strict
Nanotechnology Forum. In addition to this, definition of nanotechnology has allowed the
laboratories around the world are working research emphasis to broaden, encompassing
on new approaches and on new ways to scale many areas of work that have traditionally
up nanotechnology to industrial levels. For been referred to as chemistry or biology
example, the first factories to manufacture (DTI, 2002). Thus, the first major
carbon nanotubes and fullerenes are under characteristic of activity grouped under this
construction in Japan (DTI, 2002). section is that contemporary R&D cuts
across a wide range of industrial sectors.
In spite of these developments, there has In some cases, major markets are fairly well
been criticism recently over the amount defined. The food industry serves as a good
of hype and, consequently, funding that example here, where there are significant
research into nanoscience and drivers at work (pers. comm., Abid Khan,
nanotechnology has received. For example, London Centre for Nanotechnology, 6 Nov
the much-heralded US National 2002). To illustrate, ‘smart’ wrappings for
Nanotechnology Initiative (NNI) has been the food industry (that indicate freshness or
criticised for using ‘nano’ as a convenient tag otherwise) are close to the market (Saxl,
to attract funding for a whole range of new 2000). By 2006, beer packaging is
science and technologies (e.g. see Roy, 2002). anticipated by industry to use the highest
This reinvention is one way of attracting weight of nano-strengthened material, at
more money because politicians like to feel 3 million lbs., followed by meats and
they are putting money into something new carbonated soft drinks. By 2011, meanwhile,
and exciting (pers. comm., Gareth Parry, the total figure might reach almost 100
Imperial College London, 22 Nov 2002). million lbs. (nanotechweb.org, 2002). In
For these reasons, the nanotechnology sector other cases, important applications are
is far broader than you would usually expect identified but the eventual market impacts
to see and the resulting lack of a clear are more difficult to predict. For example,
definition is hampering meaningful nanotechnology is anticipated to yield
discussion of its potential costs or benefits. significant advances in catalyst technology.
Thus, if we use the standard definition given If these potential applications are realised
above, we can say that nanoscience and then the impact on society will be dramatic
technology have been around for several as catalysts, arguably the most important
decades, particularly in research, technology in our modern society, enable the
development, and manufacturing in IT. production of a wide range of materials and
Rather, it is the wide availability of tools and fuels (Saxl, 2000).
information to diverse scientific communities
that has generated the current interest in this A second characteristic of current work in
area (Chaudhari, 2001). this area is that the kinds of materials and
processes being developed are necessarily
‘technology pushed’: urged on by the
potential impacts of nanotechnology, the
R&D community is achieving rapid advances
in basic science and technology. This level of
scientific interest is gauged by Compano and
Hullman (2001) who examine the world-
14
Table 1: Summary of the major nanomaterials currently in research and development and their potential applications.
Clusters of atoms
Quantum wells Ultra-thin layers – usually a few nanometres thick – CD players have made use of quantum Current – 5 years
of semiconductor material (the well) grown between well lasers for several years. More
barrier material by modern crystal growth technologies recent developments promise to make
(Saxl, 2000). The barrier materials trap electrons in the these nanodevices commonplace in
ultra-thin layers, thus producing a number of useful low-cost telecommunications and optics.
properties. These properties have led, for example, to
the development of highly efficient laser devices.
Quantum dots Fluorescent nanoparticles that are invisible until ‘lit up’ Telecommunications, optics. 7–8 years
by ultraviolet light. They can be made to exhibit a range
of colours, depending on their composition
(Miles and Jarvis, 2001).
Polymers Organic-based materials that emit light when an electric Computing, energy conversion. ?
current is applied to them and vice versa
(pers. comm., Jenny Nelson, Imperial College London,
2 Dec 2002).
Nanocapsules Buckminsterfullerenes are the most well known Many applications envisaged Current – 2 years
example. Discovered in 1985, these C60 particles are e.g. nanoparticulate dry lubricant
1nm in width. for engineering (Saxl, 2000).
Catalytic nanoparticles In the range of 1–10 nm, such materials were Wide range of applications, including Current – ?
in existence long before it was realised that they materials, fuel and food production,
belonged to the realms of nanotechnology. health and agriculture (Hay and
However, recent developments are enabling a given Shaw, 2000).
mass of catalyst to present more surface area for
reaction, hence improving its performance (Hay and
Shaw, 2000). Following this, such catalytic nanoparticles
can often be regenerated for further use.
Carbon nanotubes Two types of nanotube exist: the single-wall carbon Many applications are envisaged: space Current – 5 years
nanotubes, the so-called ‘Buckytubes’, and multilayer and aircraft manufacture, automobiles
carbon nanotubes (Hay and Shaw, 2000). Both consist and construction. Multi-layered
of graphitic carbon and typically have an internal carbon nanotubes are already available
diameter of 5 nm and an external diameter of 10 nm. in practical commercial quantities.
Described as the ‘most important material in Buckytubes some way off large-scale
nanotechnology today’ (Holister, 2002), it has been commercial production (Saxl, 2000).
calculated that nanotube-based material has the potential
to become 50–100 times stronger than steel at one sixth
of the weight.
Self-assembling Organic or inorganic substances spontaneously form A wide range of applications, based 2–5 years
monolayers (SAMs) a layer one molecule thick on a surface. Additional on properties ranging from being
layers can be added, leading to laminates where each chemically active to being wear
layer is just one molecule in depth (Holister, 2002). resistant (Saxl, 2000).
Nanoparticulate Coating technology is now being strongly influenced Sensors, reaction beds, liquid crystal 5–15 years
coatings by nanotechnology. E.g. metallic stainless steel manufacturing, molecular wires,
coatings sprayed using nanocrystalline powders lubrication and protective layers, anti-
have been shown to possess increased hardness corrosion coatings, tougher and harder
when compared with conventional coatings (Saxl, 2000). cutting tools (Holister, 2002).
Nanostructured materials
Nanocomposites Composites are combinations of metals, ceramics, A number of applications, particularly Current – 2 years
polymers and biological materials that allow multi- where purity and electrical conductivity
functional behaviour (Anton et al., 2001). When characteristics are important, such as
materials are introduced that exist at the nanolevel, in microelectronics. Commercial
nanocomposites are formed (Hay and Shaw, 2000), exploitation of these materials is
and the material’s properties – e.g. hardness, currently small, the most ubiquitous
transparency, porosity – are altered. of these being carbon black, which finds
widespread industrial application,
particularly in vehicle tyres
(Hay and Shaw, 2000).
Textiles Incorporation of nanoparticles and capsules into Military, lifestyle. 3-5 years
clothing leading to increased lightness and durability,
and ‘smart’ fabrics (that change their physical
properties according to the wearer’s clothing)
(Holster, 2002).
various atmospheric dangers, such as techniques that will allow the large-scale
lightning strikes, micrometeors, and human- economic production of nanotechnology
made space debris. products, and the necessary support for
quality control (DTI, 2002). Because of the
The market impetus behind such essential nature of this category, its influence is
developments, then, is clear: the conventional far greater than is reflected in the size of the
space industry is anticipated as the first economic sectors producing these products.
major customer, followed by aircraft For this reason, the tools and techniques
manufacturers. However, as production costs highlighted below have a strong commercial
drop (currently US$20–1200/g), nanotubes future and the greatest number of established
are expected to find widespread application companies (pers. comm., Gareth Parry,
in such large industries as automobiles and Imperial College London, 22 Nov 2002). The
construction. In fact, it is possible to following sections cover methods for top-
conceive of a market in any area of industry down and bottom-up manufacture, software
that will benefit from lighter and stronger modelling and nanometrology. However, in
materials (Holister, 2002). It is expectations the near future, this area will mainly feature
such as these that are currently fuelling the extensions of conventional instrumentation
race to develop techniques of nanotube mass- and top-down manufacturing. More futuristic
production in economic quantities. The ETC molecular scale assembly remains distant
Group (2002b) states that there are currently (Miles and Jarvis, 2001).
at least 55 companies involved in nanotube
fabrication and that production levels will 2.2.4.1 Top-down manufacture
soon be reaching 1 kg/day in some Scanning Probe Microscope. This is the
companies. For example, Japan’s Mitsui and general term for a range of instruments with
Co. will start building a facility in April 2003 specific functions. Fundamentally, a
with an annual production capacity of nanoscopic probe is maintained at a constant
120 tons of carbon nanotubes (Fried, 2002). height over the bed of atoms. This probe can
The company plans to market the product to be positioned so close to individual atoms
automakers, resin makers and battery that the electrons of the probe-tip and atom
makers. In fact, the industry has grown so begin to interact. These interactions can be
quickly that Holister (2002) believes that the strong enough to ‘lift’ the atom and move it
number of nanotube suppliers already in to another place (pers. comm., Gareth Parry,
existence are not likely to be supported by Imperial College London, 22 Nov 2002).
available applications in the years to come.
Fried (2002) also supports this contention, Optical Techniques. These techniques can be
stating that the ‘carbon nanotube field is used to detect movement – obviously
already over-saturated’. important in hi-tech precision engineering.
Optical techniques are, in theory, restricted in
2.2.4 Tools and fabrication resolution to half the wavelength of light
It is a simple statement of fact that in order to being used, which keeps them out of the lower
make things you must first have the nanoscale, but various approaches are now
fabrication tools available. Therefore, many of overcoming this limitation (Holister, 2002).
the nanomaterials covered above are co-
evolving with a number of enabling Lithographics. Lithography is the means by
technologies and techniques. These tools which patterns are delineated on silicon chips
provide the instrumentation needed to and micro-electrical-mechanical systems
examine and characterize devices and effects (MEMS). Most significantly, optical
during the R&D phase, the manufacturing lithography is the dominant exposure tool in
16
use today in the semiconductor industry’s Similarly, the latter area of devices range
Complementary Metal Oxide Semiconductor from the production of new chemical and gas
(CMOS) process sensors, optical sensors, solar panels and
other energy conversion devices, to bio-
2.2.4.2 Bottom-up manufacture implants and in vivo monitoring. The basis
The tools here support rather more futuristic of these technologies is an organic film (the
approaches to large-scale production and responsive layer) which can be deposited on
nanofabrication based on bottom-up a hard, active electronic chip substrate. The
approaches, such as nanomachine production solid-state chip receives signals from the
lines (Miles and Jarvis, 2001). This approach organic over-layer as it reacts to changes in
is equivalent to building a car engine up from its environment, and processes them. The
individual components, rather than the less applications for these new materials and
intuitive method of machining a system devices are summarised in Table 2.
down from large blocks of material. Indeed,
Table 2: Applications for new materials and devices
although such techniques are still in their
resulting from self-assembly and self-organisation.
infancy, the DTI (2002) report a recent
Name Technique Application
movement away from top-down techniques
towards self-assembly within the New materials
international research community. Scientists Sol-gel technology Inorganic and The design of different
(Miles and Jarvis, 2001) organic component types of materials;
and engineers are becoming increasingly able
combination. functional coatings.
to understand, intervene and rearrange the
Intercalation of Intercalation of Toxicity testing, drug
atomic and molecular structure of matter, polymers (Miles and polymers with other delivery and drug
and control its form in order to achieve Jarvis, 2001) materials (DNA, drugs). performance analysis.
specific aims (Saxl, 2000). Nano-emulsions Nanoparticle size and Production of required
(Saxl, 2000) composition selected. viscosity and absorption
Self-assembly and self-organisation. Self- characteristics.
assembly refers to the tendency of some Biomimetics Design of systems, High strength, structural
(Anton et al., 2001) materials and their applications, such as
materials to organise themselves into ordered
functionality to mimic artificial bones and
arrays (Anton et al., 2001). This technique nature. teeth.
potentially offers huge economies, and is
New devices
considered to have great potential in
Field-sensing devices Combination of Biosensing and
nanoelectronics. In particular, the study of (Saxl, 2000) molecular films with optical switching.
the self-assembly nature of molecules is optical waveguides
proving to be the foundation of rapid growth and resonators.
in applications in science and technology. For Material-sensing Surfaces of liquid Gas and chemical
devices (Saxl, 2000) crystals or thin sensing.
example, Saxl (2000) reports that the
membranes and other
Stranski–Krastonov methods for growing organic compounds
self-assembly quantum dots has rendered the can be used to detect
molecules via structural
lithographic approach to semiconductor
or conductive changes.
quantum dot fabrication virtually obsolete.
In addition, self-assembly is leading to the
fabrication of new materials and devices. The 2.2.4.3 Software Modelling
former area of materials consists of new Molecular modelling software is another
types of nanocomposites or organic/inorganic fabrication technique of wide-ranging
hybrid structures that are created by applicability as it permits the efficient
depositing or attaching organic molecules to analysis of large molecular structures and
ultra-small particles or ultra-thin manmade substrates (Miles and Jarvis, 2001). Hence,
layered structures (Hay and Shaw, 2000). it is much used by molecular
are to operate correctly, for example X-ray Japan 120 135 157 245 465 ~650 NA
optical components and mirrors used in laser Others** 70 83 96 110 380 ~520 NA
technologies (Saxl, 2000). Total 432 559 687 825 1502 2174 NA
18
regarded as the benchmark against which elaborating a ‘conceptual template for
nanotechnology funding should be compared achieving new levels of war-fighting
(Roman, 2002). Indeed, Howard (2002) effectiveness’ (DoD, 2002). This table
states that, ‘while other governments are provides a fairly accurate picture of current
investing in a range of nanotechnology research priorities in the US. However, state
research, the US effort is by far the most funding, which can sometimes be substantial,
substantial.’ From 1985–1997 the total is not included in the estimates. For example,
support for projects related to the state of California, which is home to
nanotechnology was estimated at US$452 virtually all the work in molecular
million, coming in roughly equal parts from nanotechnology, has invested US$100 million
the NSF, various industrial sponsorship, and in the creation of a California Nanosystems
other government funding. Then in 2000, the Institute. And neither are the figures static;
much-heralded NNI was launched – a multi- levels of funding are anticipated to increase
agency programme designed to provide a big rapidly once the economic benefits of US
funding boost for nanotechnology. There are funding begin to be felt, whether in new
currently 10 US government partners in the company start-up activity, or progress
NNI3. These are shown in Table 4. towards military or social goals.
Table 4: Breakdown of spending on the US’s National
Nanotechnology Initiative from 2001–2003 (US$million). 2.2.5.2 Far East
Recipient 2001 2002 2003 Table 5 shows the levels of 2002 government
actual estimate proposed spending on nanotechnology within five
National Science countries in the Far East. On average, these
Foundation 145 199 221 figures are lower than in the US although,
Department of Defence 125 180 201 given the increased purchasing power in
Department of Energy 78 91 139 countries such as China, they may be
considered as relatively high (Roman, 2002).
National Aeronautics 0 46 49
and Space Administration However, while the figures given are up-to-
National Institute 40 41 43 date, the time-scales over which they operate
of Health are ambiguous.
National Institute of 28 37 44 Table 5: Top five government spending
Standards and Technology on nanotechnology in the Far East in 2002
(US$million).
Environmental 5 5 5
Protection Agency Country Spending
Department of 0 2 2 Japan 750
Transportation
China 200
US Department 0 2 5
of Agriculture Korea 150
Roman, 2002.
Table 4 shows that the NSF and Department
of Defence (DoD) are the two major Of all the countries shown in Table 5,
recipients of investment in nanoscience and Japan’s nanotech investments are by far the
technology R&D. Indeed, the NSF has greatest. Indeed, it is universally agreed that
designated ‘nanoscale science and Japan has the only fully co-ordinated and
engineering’ as one of its six priority areas, funded national policy of nanotechnology
while the DoD has dedicated its funding to research. The most prominent product of this
20
2.3 Applications and Markets (DTI, 2002). In spite of these difficulties,
some forecasts exist that do hint at the kind
2.3.1 Introduction of growth we might expect.
The applications of nanotechnology are Table 8: Summary of future estimated
extremely diverse, mainly because the field is global markets in nanotechnology.
interdisciplinary (Miles and Jarvis, 2001). In Year Estimated global market
addition, the effect that nanotechnology will
2001 £31–55 billion
have during the next decade is difficult to
2005 £105 billion
estimate because of potentially new and
unanticipated applications. For example, if 2008 £500 billion
22
markets around 70% (DTI, 2002). A number the force behind Moore’s law, which predicts
of recent forecasts, although varying greatly, that the processing power of ICs will double
reflect this market confidence. For example, every 18 months (Glinos, 1999). Based on
Miles and Jarvis (2001) put the market for Moore’s law, industry predictions are
nanotechnology-based IT and electronics summarised in Table 9.
devices at around US$70 billion by 2010. A
second estimate states that nanotechnology Semiconductor industry associations assume
will yield an annual production of about that they will be close to introducing 100 nm
US$300 billion for the semiconductor ground-rule technology by 2004 (Compano,
industry and about the same amount again 2001). The significance of this lies in the fact
for global integrated circuits sales within that 100 nm is widely viewed as a kind of
10–15 years (NSF, 2001). Similarly, for ‘turning point’, where many radically new
micro- and nanotechnology systems in the technologies will have to be developed. To
telecommunications sector, the market is begin with, optical lithography will become
presently estimated as being in the order of obsolete somewhere around 100 nm. As a
US$35 billion with an anticipated compound result, ‘next generation lithography’ options
annual growth rate of around 70%. are currently being investigated. These are
summarised in Table 10.
2.3.2.1 Moore’s Law
Table 10: Maturity of lithography options.
The microelectronics industry had looked
ahead and seen serious challenges for its Year of introduction 2001 2003 2006 2009
basic CMOS process. CMOS technology has Minimum feature size 150 120 90 65
been refined for over 20 years, driving the Optical 193 nm X* X
‘line width’-the width of the smallest feature Optical 157 nm X X
in an integrated circuit (IC)-from 10 mm
Extreme UV X X
down to 0.25 µm (Doering, 2001). This is
X-rays X
Table 9: Anticipated technological
Electron beam X X
computing developments for 2001–2014.
Ion beam X X
Feature Year
Printing X
2001 2003 2005 2008 2011 2014
*An ‘X’ designates the date at which the respective
Memory
fabrication technology is expected to become economically
Minimum feature 150 120 100 70 50 35 viable for mass production.
size DRAM
Adapted from Compano, 2001.
(1/2 pitch in nm)
Gbits/chip 2 4 8 24 68 194
Excluding the printing process, each
Density 0.49 0.89 1.63 4.03 9.94 24.50 fabrication technique essentially works on
(Gbits/cm2)
the same principle where a reactive silicon-
Logic (processing power) based agent is exposed to increasingly
Minimum feature 100 80 65 45 30–32 20–22 focused electromagnetic radiation: optical to
size (gate length X-rays representing a successive reduction in
in nm)
photon wavelength; E-beam and ion beam
Density (million 13 24 44 109 269 664
projection technologies using focused
transistors per cm2)
electron and ion beams respectively. All of
Logic clock (GHz) 1.7 2.5 3.5 6.0 10.0 13.5
these techniques are currently under active
DRAM: Dynamic Random Access Memory, evaluation-the aim is to have the appropriate
a type of memory used in most personal computers.
equipment for the corresponding time-frame.
Adapted from Compano, 2001
To date, X-ray and ion bean projection have
24
with conventional electronic computers matter, however, because of the greater
(Anton et al., 2001). complexity involved-their applicability
remains limited until total control over the
QIP. This crosses the disciplines of quantum emerging structures in terms of wiring and
physics, computer science, information their interconnections can be obtained. These
theory and engineering with the aim of are formidable obstacles. Self-assembly,
harnessing the fundamental laws of quantum therefore, will likely be combined initially
physics to ‘dramatically improve the with some more traditional top-down
acquisition, transmission and processing of approaches. For example, many believe that
information’ (Miles and Jarvis, 2001). QIP inducing molecular components to self-
represents computing at the smallest possible assemble on a patterned substrate in some
scale, in which one atom is equivalent to one sort of hybrid system will represent the first
byte of information. Other aspects of commercialisation of nanoelectronics
quantum computing also considered (Holister, 2002).
attractive relate to their massive parallelism
in computation (i.e. the ability to perform 2.3.2.3 Summary of applications
simultaneous calculations) (Holister, 2002). The main drivers for current applied
These concepts are qualitatively different microelectronics research are computing,
from those employed in traditional telecommunications, consumer electronics
computers and will hence require new and military applications. It is not evident
computer architectures. A preliminary survey how long personal computing will act as a
of work in this area by Anton et al., (2001) driver. On the one hand, personal computers
indicates that quantum switches are unlikely (PCs) already offer sufficiently good
to overcome major technical obstacles, such performance for a large number of users; on
as ‘error correction, de-coherence and signal the other, new applications, such as
input/output’ within the next 15 years. If this automatic voice recognition or PC wireless
proves to be the case, QIP-based computing, communications, may give further impulses
as with molecular nanoelectronics, does not for further progress (Compano, 2001).
appear to be competitive with traditional Military applications have a restricted
digital electronic computers for some time. volume but are of strategic importance.
Thus, it is generally anticipated that most
Computational self-assembly. A major barrier new technologies within this area will emerge
to the introduction of nanoelectronics is that in (US) military use first before eventually
there are no established mass production finding their way into the civil sector (Saxl,
techniques for creating devices on a 2000). These and other ideas are summarised
commercial basis (Saxl, 2000). In the short in Table 11.
to medium term, Table 10 covers the most
promising fabrication approaches. In the 2.3.3 Pharmaceuticals and medicine
long-term, however, more ambitious bottom- Nanotechnology, combined with
up methods based on self-assembly biotechnology, are the underpinning
techniques are proposed. Bottom-up technologies pushing the rapid advances in
approaches are elegant, cheap and possibly ‘genomics, combinatorial chemistry, high
enormously powerful techniques for future throughput robotic screening, drug discovery,
mass replication. The relatively gene sequencing and bioinformatics and their
straightforward architecture of molecular applications’ (Saxl, 2000). In medicine,
memory means that self-assembly techniques advances can take place at the nanoscale
in this area may bear fruit in a few years where, for example, either passive or active
(Table 11). Tackling processors is another nano-engineered systems can be used that
Pre 2015
Photonic crystal technologies Optical communication sector, i.e. fibre Still in basic R&D, but very
(Miles and Jarvis, 2001). optics. Photonic integrated circuits can strong commercial interest
be nearly a million times denser than emerging.
electronic ones. Their tighter
confinement and novel dispersion
properties also open up opportunities
for very low power devices
Spintronics – the utilisation Ultra-high capacity disk drives and A read head has been
of electron spin for significantly computer memories. demonstrated that can deal with
enhanced or fundamentally storage densities of a terabit per
new device functionality square inch. In 2001 Fuji
(Science Blog, 2002). announced a new magnetic
coating promising 3-gigabyte
floppy disk.
Post 2015
Molecular nanoelectronics Circuits based on single molecule and Single atom transistor
(including DNA computing) single electron transistors will appear, demonstrated recently. Still
(Compano, 2001). initially in special applications. immature, but huge potential
(Miles and Jarvis, 2001).
Quantum information Several researchers have devised Still in pure research phase,
processing (QIP) algorithms for problems that are very although some US defence
(Compano, 2001). computationally intensive (and thus money has been made available
time-consuming) for existing digital (Holister, 2002).
computers, which could be made much
faster using the physics of quantum
computers. E.g. factoring large numbers
(essential for cryptographic applications),
searching large databases, pattern
matching, simulation of molecular
and quantum phenomena
(Anton et al., 2001).
26
enable the required dose of drug to be with no guarantee of success. In response,
delivered at the correct time to the target some companies are trying to hurry the long
area, or at the macro-level, such as induced clinical phase required in Western medicine
tissue growth. This reduces unwanted side (Ho, 2002a). However, powerful incentives
effects, improves patient compliance, leads to remain to investigate new techniques that can
lower doses and opens up new possibilities more effectively deliver or target existing
(that would be impossible without drugs (Saxl, 2000). In addition, many of these
nanotechnology approaches) (Miles and new tools will have foundation in current
Jarvis, 2001). The size of this market is the techniques: a targeted molecule may simply
main driving force behind such innovation. add spatial or temporal resolution to an
LaVan and Langer (2001) predict that: existing assay. Thus, although many potential
‘fundamental changes in drug production applications are envisaged, the actual near-
and delivery are expected to affect about half future products are not much more than
of the [US]$380 billion world- wide drug better research tools or aids to diagnosis (Ho,
production in the next decade.’ At present, 2002a). These are summarised in Table 12.
nanotechnology is estimated to have a 1%
stake in this, but whole sectors will continue 2.3.4 Energy
to grow and this contribution is expected to The global energy sector is likely to be
increase rapidly (Ho, 2002a). The US is particularly affected by coming advances in
widely recognised as a leader in this area, nanotechnology. To illustrate, significant
with a company market share of about 40%, changes in lighting technologies are expected
and many applications close to the market. in the next 10–15 years. Semiconductors used
in the preparation of light-emitting diodes can
2.3.3.1 Drug delivery increasingly be sculpted on nanoscale
The most promising aspect of dimensions. Projections indicate that such
pharmaceuticals and medicine as it relates to nanotechnology-based advances have the
nanotechnology is currently drug delivery. In potential to reduce world-wide consumption
the words of LaVan and Langer (2001): ‘It is of energy by more than 10% (NSF, 2001).
likely that the pharmaceutical industry will The various applications showing most
transition from a paradigm of drug discovery promise are summarised in Table 13.
by screening compounds to the purposeful
engineering of targeted molecules.’ At present, Most current photovoltaic (PV) production is
there are 30 main drug delivery products on based upon crystalline and amorphous
the market. The total annual income for all of silicon technologies. However, as Table 13
these is approximately US$33 billion with an shows, research is now focusing upon new
annual growth of 15% (based on global technologies which may result in significant
product revenue) (Miles and Jarvis, 2001). reductions in PV costs and/or improvements
Two major drivers are primarily responsible in efficiency. Nanotechnology is anticipated
for this increase in the market. First, present to play an important part in this future.
advances in diagnostic technology appear to Although total PV power output remains
be outpacing advances in new therapeutic relatively low, the industry is growing
agents. Highly detailed information from a rapidly-the production of PV modules
patient is becoming available, thus promoting expanded by 40% in 1997 (Saxl, 2000). This
much more specific use of pharmaceuticals increase is largely due to the building and
(LaVan and Langer, 2001). Second, the construction industry, the largest and fastest
acceptance of new drug formulations is growing sector at present. In addition,
expensive and slow, taking up to 15 years to developing countries represent a potentially
obtain accreditation of new drug formulas vast market (pers. comm., Jenny Nelson,
Diagnostics
Drug delivery
Nanoporous materials Evading body’s immune When coupled to sensors, Pre-clinical: an insulin-
(Holister, 2002). system whilst directing a drug-delivering implants delivery system is being
therapeutic agent to the could be developed. tested in mice.
desired site.
‘Pharmacy-on-a-chip’ Monitor conditions and act E.g. Diabetes treatment. More distant than ‘lab-on-
(Saxl, 2000). as an artificial means of a-chip’ technologies.
regulating and maintaining
the body’s own hormonal
balance.
Cellular manipulation Manipulation and coercion Persuasion of lost nerve More distant: 5–7 years.
(Miles and Jarvis, 2001). of cellular systems. tissue to grow; growth
of body parts.
28
Table 13: Summary of applications for energy processing.
Polymer materials Solar cells (pers. comm., Jenny Nelson, The research stage has
(organic). Imperial College London, 2 Dec 2002). advanced much more quickly
Current developments aim to balance moderate than expected. As a result,
efficiency with low cost. Another big advantage polymer-based PV cells should
is that these layers can easily be incorporated enter the market in 5 years.
into appliances. Current problems stem from
the material’s instability.
Combinations of Dye-sensitised solar cells made from a thin Low power applications will
organic and hybrid layer (pers. comm., Jenny Nelson, enter market first. Limited
inorganic molecules. Imperial College London, 2 Dec 2002). These commercialisation already
cells are potentially very cheap because occurring (e.g. by Sustainable
fabrication is from cheap, low purity materials Technologies International).
by simple and low cost procedures (Saxl, 2000).
Photocatalytic water treatment.
Quantum wells Quantum-well solar cells (pers. comm., Jenny Pure research.
(inorganic). Nelson, Imperial College London, 2 Dec 2002).
Current research is taking place in high-efficiency
applications because the infrared part of the
solar spectrum may be absorbed.
Fuel conversion/storage
30
goal of this research centre is to greatly Table 14: US historical funding for technology
enhance the protection and survival of the transitioning into the marketplace.
32
surrounding nanotechnology has, to some attracted a great deal of public interest, and
extent, damaged investment potential (DTI, impressive demonstrations have been made
2002). For example, Schultz (2002) advocates of microscopic devices. For example, in
the need for nanotech researchers and August 2001 scientists from Osaka
supporters to dampen unquestioning University built the smallest micromechanical
enthusiasm for nanotechnology. This is system ever, a spring whose arm is only
because, without discussion of the potential 0.3 µm wide (Ho, 2002a). However,
pitfalls, future nanotechnology research could although almost qualifying as a nanodevice,
be subjected to such extreme pressure that the question of whether it is possible to
funding is jeopardised and research progress attain extreme capability and, if so, how to
is slowed, perhaps halted altogether in some develop the field, is a point of contention in
cases. This realisation has quickly lead to an both scientific and policy circles (Nelson and
attempt by industry to diffuse some of the Shipbaugh, 1995).
wilder claims surrounding the field. Glenn
Fishbine’s Investor’s Guide to Nanotech and In spite of the above controversies, it remains
Micromachines (2002) provides one such clear that bottom-up technologies, while
example: it is through this type of work that having the potential to be immensely
the science fiction aspects of the debate are important in the longer term, are not likely
now receding (pers. comm., Abid Khan, in the near future (DTI, 2002). However,
London Centre for Nanotechnology, 6 Nov some products benefiting from research into
2002). However, in spite of these molecular manufacturing may be developed
developments, it is clear that the distinction in the near term. As initial nanomachining,
between near-future applications of novel chemistry and protein engineering
nanotechnology (see Section 2.3) and some (or other biotechnologies) are refined, initial
of the more visionary aspects of the debate products will likely focus on those that
has become blurred. An attempt to substitute for existing high-cost, lower-
distinguish between the two areas here will efficiency products. Likely candidates for
help draw out some of the more legitimate these technologies include a wide variety of
concerns currently being voiced about sensor applications, tailored biomedical
nanotechnology in the following section. products (including diagnostics and
therapeutics), extremely capable computing
2.4.2 Molecular nanotechnology and storage products, and unique, tailored
The more hyped aspects of nanotechnology materials (i.e. smart materials using
have generally revolved around MNT. nanoscale sensors, actuators, and perhaps
Proponents of this approach suggest that controller elements) for aerospace or similar
environmentally clean, inexpensive, and high-capability needs (Nelson and
efficient manufacturing of structures, devices, Shipbaugh, 1995). Predictions of when
and ‘smart’ products, based on the flexible bottom-up processes will begin to become
control of architectures and processes at an available on a widespread basis vary across
atomic or molecular scale of precision, may the literature. In general, the hyped aspects
be feasible in the near future (i.e. 10–20 of the industry are operating around a
years from the present). The ambitious goal 20-year time-scale, with estimates for
is to produce complex products on demand economically viable self-assembly techniques
using simple raw materials, such as by tending to convene around 2015 (Ho,
inserting the basic chemical elements in a 2002a). However, to reach a fully mature
molecular assembly factory to yield a nanotechnology society – where it is possible
common household appliance (Nelson and to manipulate objects on all scales from atom
Shipbaugh, 1995). These visions have to macroscopic – is expected to take at least
34
companies, Zyvex (2002), aims ‘to become Furthermore, other commentators such as
the leading world-wide supplier of tools, Ho (2002c), point to major problems
products, and services that enable adaptable, concerning energy sources and dissipation, or
affordable, and molecularly precise just the sheer complexity of the task at hand.
manufacturing’ and offers a ‘variety of For example, diamond assemblies might be
products, services, and licensing relatively easy to assemble; other structures,
opportunities,’ including a number of such a biological configurations, are
nanomanipulation devices. Such nano- infinitely more complicated.
advocates claim the first major breakthrough
in this area might occur as early as 2007. 2.5 Concerns
36
2.5.2.2 Self-replication developing nanofabrication techniques for
Self-replication is probably the earliest- manufacturing nanoelectronic devices in huge
recognised and best-known long-term danger volumes at very low cost, then the impact on
of MNT. This centres upon the idea that self- society will be enormous. The potentially
replicating nanorobots capable of functioning disruptive nature of nanotechnology has
autonomously in the natural environment already been highlighted in earlier sections
could quickly convert that natural through its ability to generate major new
environment (i.e. ‘biomass’) into replicas of paradigm shifts in how things are generated,
themselves (i.e. ‘nanomass’) on a global such as a shift from top-down to bottom-up
basis. Such a scenario is usually referred to manufacturing techniques. This section further
as the ‘grey goo’ problem but perhaps more elaborates upon this and similar concerns.
properly termed ‘global ecophagy’ (Freitas,
2000). The main feature that distinguishes 2.5.3.1 Medical ethics
runaway replication as a long-term The ethical questions that have been raised in
environmental concern is the extreme recent years following the advancement of
difficulty involved in constructing machines such technologies as gene therapy are similar
with the adaptability of living organisms. to in scope and philosophy to nanotechnology.
As Freitas (2000) notes: For example, the emergence of highly specific
drug therapies, a nanobased technique that
‘The replicators easiest to build will be features prominently in earlier sections of this
inflexible machines, like automobiles or report, may result in genetic discrimination.
industrial robots…To build a runaway That is, discrimination directed against an
replicator that could operate in the wild would individual or family based solely on an
be like building a car that could go off-road apparent or perceived genetic variation from
and fuel itself from tree sap. With enough the ‘normal’ human genotype (LaVan and
work, this should be possible, but it will hardly Langer, 2001). The major concern here lies in
happen by accident. Without replication, the end result of going down such a road: that
accidents would be like those of industry the de-selection of characteristics judged
today: locally harmful, but not catastrophic to unwanted by societies (referred to as negative
the biosphere. Catastrophic problems seem eugenics) will be viewed as the right,
more likely to arise though deliberate misuse, responsible, moral thing to do, as will cures
such as the use of nanotechnology for military and enhancements (Wolbring, 2002).
aggression’ (see below). Similarly, on a longer time-scale, concerns
over nanotech applications for enhancing the
This is not to imply, however, that the risk performance of the human body might also
that molecular machines designed for arise. A major question here is whether such
economic purposes might replicate enhancements can be forced upon people,
unchecked and destroy the world should be either when in a position to make a decision
written off altogether: while the danger for themselves or, more controversially,
seems slight, even a slight risk of such a against their will.
catastrophe is best avoided (Zyvex, 2002).
To this end, David Forrest (1989) has 2.5.3.2 The nano-divide
produced a set of guidelines to assure that If Moore’s law holds and the miniaturisation
molecular machines and their products are of PCs continues unchecked well into the
developed in a safe and responsible manner. 21st Century, then it seems likely that, in the
long-term, society will get to a point where
2.5.3 Socio-political concerns people can carry computers 24 hours a day.
Clearly, if scientists are successful in As Chaudhari (2001) states: ‘We are evolving
38
sketch out the underlying rationale for such House Fact Sheet (2000) promised that the
an occurrence: impact nanotechnology has on society from
legal, ethical, social, economic, and workforce
‘It is clear that offensive weapons made using preparation perspectives would be studied.
advanced nanotechnology can only be These aims have already been realised to
stopped by defensive systems made using some extent. For example, the 2001 NSF
advanced nanotechnology as well. If one side report entitled Implications of Nanoscience
has such weapons and the other doesn’t, the and Nanotechnology takes a long, hard look
outcome will be swift and very lopsided. This at a range of hypothetical social ramifications
is just a specific instance of the general rule (Roco and Bainbridge, 2001).
that technological superiority plays an
important and often critical role in This industry strategy has been received with
determining the victor in battle. Clearly, we mixed reaction. Some commentators, such as
will need much further research into Ho (2002a), have praised scientists for
defensive systems as this technology becomes informing the public with ‘clarity and
more mature.’ candour.’ Others, on the other hand have not
been nearly so generous in their assessment.
2.5.4 Public acceptance of nanotechnology Herrera (2002), for example, sums up the
In spite of the concerns highlighted above, present state of the nanotech industry as
both precautionary principle and industry being comparable to a ‘sitting duck’, just as
advocates agree that there is time to create biotech was during the 1990s, because it is
dialogue and consensus that could prevent the not taking the issue of public acceptance
kind of confrontations occurring that plagued seriously. Herrera continues: ‘Ask members of
the development of biotechnology. In this the nanotechnology community if there are
way, the objective of industry is to launch any obvious or potential controversies that
pre-emptive strikes against any problems with they should be watching for, and they will say
public acceptance of nanotechnology that ‘no’… Scientists think about ethics but they
might arise down the line (Gorman, 2002). don’t let it interfere with their work.’
The earliest example of this is the Foresight
Institute, a think-tank founded in 1986 At present, the majority of controversy in
primarily to facilitate public understanding this area surrounds the interaction of
and discussion of the policy issues nanomaterials with the environment and
surrounding the development and deployment their implications for human health. Vicki
of nanotechnology. More recently, nanotech Colvin of CBEN believes that ‘scientists’
researchers have been urged to build on the experience with other particulate matter
example of the Ethical and Social argues for a thorough examination of how
Implications (ELSI) project (an nanoparticles might react in mammalian
interdisciplinary effort within the Human systems when they are inhaled or when there
Genome Project). That is, to ‘take a hard is skin exposure’ (cited in Schultz, 2002). In
look at potential ethical and cultural issues, addition, nanotech manufacturing processes
but follow through much more carefully and need to be examined for potential health
get out ahead of the public’ (Paul Thompson, impacts, for example the solvents used in the
Professor of Ethics at Purdue University, gases produced in the manufacture of carbon
quoted in Leo, 2001). Indeed, the NNI has nanotubes. Outside of manufacturing,
long acknowledged a need to integrate researchers should investigate the possible
societal studies and dialogues concerning the consequences of nanoparticles entering and
perceived dangers of nanotechnology with its accumulating in the food chain. Indeed, some
investment strategy, and the resulting White of the ongoing work by CBEN, and other
40
important issues to be identified earlier, the we are unlikely to witness any radical
right questions to be raised, and necessary developments during the next 15 years unless
corrective actions to be taken. It does seem a series of fundamental breakthroughs occur
likely that some form of regulatory control between now and then. However, as the
will be necessary to assure that range of associated tool and fabrication
nanotechnology is developed safely – ‘safe techniques begin to mature, the field is set to
designs, safe procedures and methods to test become increasingly commonplace in the
for potentially hazardous assemblers can be coming decades. Ultimately, then, the longer-
incorporated into standards by consensus of term structural impact of nanotechnology on
interested parties’ (Forrest, 1989). The a whole range of sectors – in manufacturing,
greatest danger, however, appears to be transport, services and domestic practice –
intentional abuse of the technology, so could be substantial in 30–50 years. These
certain aspects of development should be changes are likely to be gradual as, on the
performed in a secure environment. whole, the displacement of an old technology
by a new one tends to be both slow and
2.6 Discussion incomplete (NSF, 2001).
While many of the nanotechnologies covered
in this part of the report might appear In the meantime, a number of well-founded
advanced, it is fair to conclude that most short-term concerns remain, many of which
contemporary experimental capabilities in revolve around issues of human health.
this area are still in their infancy. This means Considering past experiences of industry
that it is extremely difficult to foresee many and government mismanagement in this area
outcomes that developments in this field will (notably through GM-related controversy),
bring over the next 10 years, let alone assess nano-advocates would do well to sit up and
their likelihood. Initially, it is probable that take note. For, although an externally
the impact of nanotechnology will be limited imposed nanotech moratorium seems both
to a few specific products and services, where unpractical and probably damaging at
consumers are willing (or able) to pay a present, industry may find such a fate
premium for new or improved performance. virtually self-imposed if they do not take
Looking further ahead, controversy surrounds the issue of public acceptance seriously. This
the possibility of realising some of the wilder report has shown some nano-advocate
visions of a nanotech-enabled future. This is awareness of environmentally-sound practice.
in spite of the fact that many of these ideas Industry must demonstrate a commitment to
stem from quite straightforward concepts this by funding the relevant research on a far
founded in solid science (Holister, 2002); greater scale than currently witnessed.
42
Lighthill Report made explicit in the UK lay, it is likely that the phenomenon developed in
less explicitly, behind a similar slow down in reaction to the kind of historical tendencies
research funding in the US (Malcolm, 2001). to oversell the industry alluded to earlier.
The next big rise in AI funding occurred in
the 1980s, mainly in reaction to Japanese 3.2 Aspects of Research
enthusiasm for the field. In Japan, the 5th
Generation project was born; the UK reacted 3.2.1 Introduction
through the Alvey initiative, which now The above section has described how
focused on ‘knowledge based systems’ so as researchers have re-evaluated their
to avoid any awkward parallels between expectations with regard to achieving strong
current research and the previously AI. Associated with this reality check is the
condemned AI. Again, both projects were recognition that classical attempts at
characterised by a lack of progress and AI modelling AI, based upon the capabilities of
research failed to make it into the digital computers to manipulate symbols, are
mainstream. Most recently, the early to mid probably not sufficient to achieve anything
1990s has seen the emergence of software resembling true intelligence. This is because
agents, and the resulting excitement has once symbolic AI systems, as they are known, are
again sparked a rise in investment. In designed and programmed rather than
addition, the field of robotics has become trained or evolved. As a consequence, they
much more influential of late, particularly function under rules and, as such, tend to be
through the entertainment industry. Many very fragile, rarely proving effective outside
of these developments are described in more of their assigned domain (Hsuing, 2002). In
detail later on in this report. other words, symbolic AI is proving to be
only as smart as the programmer who has
Today, AI is about at the same place the PC written the programmes in the first place.
industry was in 1978 (Brooks, 2001) – the
waves of enthusiasm that accompanied the In realisation of this, scientists are beginning
developments of computers have long gone to look much more closely at the mechanisms
and researchers are beginning to come to of the brain and the way it learns, evolves
terms with how hard the problems of AI and develops intelligence from a sense of
really are. However, technological know-how being conscious (Aleksander, 2002). For
is not the only obstacle that the AI industry example, AI software designers are beginning
faces – another is the purported ‘AI effect’ to team up with cognitive psychologists and
whereby the existence of AI in modern use cognitive science concepts. Another
software products go largely unnoticed example centres upon the work of the
despite the widespread use of such ‘connectionists’ who draw attention to
applications (Stottler Henke, 2002). Indeed, computer architecture, arguing that the
AI is considered by some researchers to be an arrangement of most symbolic AI
unimplementable technology: as soon as the programmes is fundamentally incapable
technology advances, the perspective shifts, of exhibiting the essential characteristics
and the quality of intelligence passes to those of intelligence to any useful degree. As an
activities that are still only in the human alternative, connectionists aim to develop AI
domain (Joseph, 2001). For example, many through artificial neural networks (ANNs).
of those in industry do not use the term Based on the structure of the nervous system,
‘artificial intelligence’ even when their these ‘computational-cognitive models’ are
company’s products rely on some AI designed to exhibit some form of learning
techniques (Stottler Henke, 2002). The exact and ‘common-sense’ by drawing links
reasons for the AI effect are uncertain, but between meanings (Hsiung, 2002). ANNs,
44
and scientific analysis of data (Stottler factory automation, military transportation
Henke, 2002). scheduling, and medical treatment planning.
These will be covered in more detail below.
3.2.3 Reasoning about plans,
programs and action 3.2.4 Logical AI
Intelligent systems must be able to plan – to This type of reasoning concerns what a
determine appropriate actions for their programme knows about the world in
perceived situation, and then execute them general, the facts of the specific situation in
and monitor the results. However, in spite of which it must act, and the goals that it must
the fact that this area has been under active accomplish (Grosz and Davis, 1994). Such
research since the 1950s, AI planning concepts are held within the programme in
applications are furthest from human-level the form of sentences of some mathematical
(Grosz and Davis, 1994). Ordinary people, logical language. The most successful
for example, manage to accomplish an example of this is an expert system, created
extraordinary number of complex tasks just when a ‘knowledge engineer’ interviews
using simple, informal thought processes experts in a certain domain and tries to
based on a large amount of common embody their knowledge in a computer
knowledge. AI, on the other hand, is far programme for carrying out some task, such
behind humans in using such reasoning as diagnosis. However, the usefulness of
except for limited jobs, and tasks that rely current expert systems also depends on their
heavily on common-sense reasoning are users demonstrating a certain level of
usually poor candidates for AI applications common-sense too.
(Stottler Henke, 2002). In the past,
researchers have mainly had to rely on the 3.2.4.1 Algorithms and genetic programming
development of algorithms that An algorithm is defined as a ‘detailed
‘automatically construct and execute sequence of actions to perform to accomplish
sequences of primitive commands in order to some task’ (FOLDOC, 2003). One branch of
achieve high-level goals’ (Weld, 1995). More algorithm theory, genetic programming, is
recently, the field of plausible reasoning has currently receiving much attention. This is a
demonstrated its feasibility in tackling the technique for getting software to solve a task
problem of representing, understanding, and by ‘mating’ random programmes and
controlling the behaviour of agents or other selecting the fittest in millions of generations.
systems in the context of incomplete or Khan (2002) elaborates: ‘Genetic algorithms
incorrect information (Weld, 1995). Another use natural selection, mutating and
development that may lead to significant crossbreeding within a pool of sub-optimal
advances in the area of artificial reasoning is scenarios. Better solutions live and worse
fuzzy logic. Traditional Western logic systems ones die – allowing the programme to
assume that things are either in one category discover the best option without trying every
or another. Yet in everyday life, we know this possible combination along the way.’
is often not precisely so. Fuzzy logic, then,
provides a way of taking into account our 3.2.5 Collaboration
common-sense knowledge that most things The ubiquity of computers, networks and
are a matter of degree when a computer is distributed information resources means that
automatically making a decision (Stottler collaboration between these entities is
Henke, 2002). Thus, in spite of the important. The field of multiagent co-
difficulties inherent in this field of AI, ordination concerns itself with the problem
planning systems have been successfully of endowing agents with the ability to
developed for several tasks to date, including communicate with each other to reach
46
the world’s economy. Obviously, in such 3.2.8.1 The US
a future scenario, the extraordinarily Historically in the US, the concept of AI
sophisticated systems used to control originated in the private sector, but the
communications, power, stock exchanges, growth of the field has depended largely on
and monetary assets can break down and public investments. Today, computer science
might come under attack. Today, given the research in the US is funded by a number of
relative complexity and unreliability of the governmental agencies. Total US government
Internet, it is not surprising that computer science research expenditures in
commentators view this scenario with 1998 were US$1399 million, with
increasing trepidation. AI, then, is seen by approximately one third devoted to what
many as having an essential role in a future was described as ‘basic research’ (Schneider
where commercial and military information and Robb, 2001). Three agencies (NSF,
warfare is a major, perhaps dominant, DARPA, and the Department of Energy
characteristic. In addition, AI is touted by [DOE]) in the US together support US$365
many (e.g. see Dertouzos, 1999) as having million of this work, and the NSF is
the potential to greatly improve human responsible for funding the lion’s share
productivity and ease of use within this (Schneider and Robb, 2001). In addition, a
prospective network. number of other agencies are of note. These
include the National Institutes of Health
As computer science, and AI in particular, (NIH) and NASA which have also pursued
is considered to be of strategic importance, AI applications of particular relevance to
it is worth here briefly examining their own separate agendas (NRC, 1999).
government funding in this area. In general,
computer science receives a relatively small Of these institutions, DARPA is credited with
proportion of the research funding in many considerable advancement of the field from
countries, even if anecdotal evidence the 1960s onwards. This hardly comes as
suggests that the fractions are increasing surprising when one considers the close link
(Schneider and Robb, 2001). This is in spite that exists between the military and
of the fact that public funding has played an computer science – in fact, the early
important part in AI research in the past, development of computing was virtually
largely because of the field’s high-risk exclusively limited to military purposes
conceptual challenges. However, the picture (Matthews, 2000b). The most famous
is complicated by the fact that international example of this concerns the development of
comparisons of research funding in this area the Internet, in which DARPA played a
are difficult to make, since different central role in the 1970s and 1980s5. More
countries use different funding methods. recently, less visible but arguably equally
France and Japan, for example, rely heavily significant developments have come to the
on national research institutes and fore. For example, a 1994 report by the
laboratories, rather than expecting most AAAI paraphrased a former director of
research to be done in university DARPA, saying that DART (the intelligent
departments. In addition, funding for AI system used for troop and material
research is reported far less thoroughly than deployment for Operation Desert Shield and
it is for nanotechnology. Consequently, Operation Desert Storm in 1990 and 1991)
relevant information has often been ‘justified DARPA’s entire investment in AI
obscure, and it has been necessary instead technology’ (cited in NRC, 1999). One
to report funding in computer science in consequence of this is that modern-day battle
general. As a rule, AI budgets will represent relies heavily on data networks. This has
a small proportion of these figures. been stressed by A. Michael Andrews, the US
48
technology for next-generation symbiotic AI are virtually endless: one measure of the
robots and systems, and new models for growth of practical applications is the
information sharing and exchanging (NII, number of patents mentioning the term AI
2002). This programme is closely linked to and related terms. According to the US
the Japanese government’s seven-year plan to Patent Office, only about 100 patents
develop humanoid robots. In fact, Japan is specifically mentioned AI a decade ago; in
the clear leader in using industrial robots as contrast, in 1999 about 1,700 patents
it accounts for over half of all units in the mentioned AI with another 3,900 or so
world (The Economist, 2001). mentioning related terms (Buchanan and
Uthurusamy, 1999). However, it is worth
The EU, too, is engaged in AI-related bearing in mind that the actual prevalence of
research. Perhaps the most ambitious of this emerging AI technology may be greater than
is related to the EU Framework VI proposal this due to classification-related difficulties
for spending €16.29 billion over 2002–2006, and the fact that such products are more
of which 27% is destined for IT (Schneider likely to be embedded in some larger system
and Robb, 2001). In addition to central EU than a stand-alone machine. In general, such
funding, individual European states are also applications are used to increase the
developing their own research agendas. In the productivity of knowledge workers by
UK, the Information Technology/Computer intelligently automating their tasks, or to
Science (IT/CS) Programme in the Engineering make technical products of all kinds easier to
and Physical Sciences Research Council use for both workers and consumers through
(EPSRC) budget for 2000/2001 is £70.3 intelligent automation of their complex
million; investment in computer science functions (Stottler Henke, 2002). It is
research is about 45% of this (EPSRC, 2003). possible now to identify four families of
Other EU countries are also of interest, intelligent systems that have broad
particularly Germany, France and applicability across a wide range of sectors
Scandinavia, where the latter is particularly (Grosz and Davis, 1994). These are
well advanced in the use of computing and intelligent simulation systems; intelligent
IT. However, other smaller countries are not information resources; intelligent project
making much in the way of substantial coaches; and robotics.
commitments to computer science research.
3.3.2 Intelligent simulation systems
3.3 Applications These applications are commonly used in a
number of different scenarios. First, an
3.3.1 Introduction Intelligent Simulation System (ISS) may be
The above section has demonstrated the generated to learn more about the behaviour
diverse and multifaceted nature of AI of an original system, when the original
research, and this work has resulted in an system is not available for manipulation. The
extensive body of principles, representations, modelling of climate systems is a good
algorithms, and spin-off technologies (Weld, example. Second, the original system may
1995). The relative state of infancy of not be available because of cost or safety
research into strong AI means that this field reasons, or it may not be built yet and the
can be put aside for the time being. Rather, purpose of learning about it is to design it
this section will attempt to elaborate upon better (Stottler Henke, 2002). Third, an ISS
weak applications of AI, where, it is fair to might be employed for training purposes in
say, considerable effort in this area has anticipation of dangerous situations, when
resulted in some real-world product success. the cost of real-world training is prohibitive.
In fact, the actual and potential uses of weak Such technologies are particularly well-
50
and collaborating in a wide range of valuable practical applications of this
design or operations teams for complex technology, developments in this area are
systems. For basic personal use, ‘interface expected to advance quickly. For example,
agents’ are computer programs that automated language translation also looks
employ AI techniques to provide active set to mature sometime between 10–15
assistance to a user during computer-based years from present.
tasks. These agents acquire their competence
by learning from the user as well as from Perhaps the most ambitious examples of
agents assisting other users. To date, AI development that are currently occurring
several prototype agents have been built in this area relate to computer learning.
using this technique (Maes, 1994). For One example is the ANN, Falcon. Designed
example, US start-ups, such as Saffron by San Diego-based HNC Software, Falcon
Technology and Manna, are marketing maintains a profile of how, when, and where
software tools that learn the individual customers use their credit cards and, from
user’s buying patterns and make this, develops an ability to discern ‘deviant’
personalised recommendations accordingly. behaviour. To date, this system is used by
nine of the ten leading US credit card
In addition to interface agents, the next companies: they claim it has improved
10 years are likely to see rapid AI fraud detection rates from 30–70% (Khan,
development occurring in speech recognition 2002). Another example – and one that is
(Hendler, 2000). Indeed, computer speech probably the most challenging in ANN
input has already arrived and is development today – is being undertaken
commercially available – many telephone by DARPA, who have launched an initiative
services use speech recognition at present. to develop a cognitive (i.e. thinking) system.
In addition, cell phones without keypads are The aim of this system is to reason in a
likely to reach the market as early as next variety of ways, learn from experience,
year. These devices are anticipated to and adapt to surprises. In the words of
enhance the use and appeal of the mobile Melymuka (2002): ‘It will be aware of its
Internet by allowing users to call up any behaviour and explain itself…It will be able
Web page from a mobile device just by to anticipate different scenarios and predict
speaking its address. Voice recognition also and plan for novel futures.’ The ultimate
has security applications: in a demo at the aim is to develop cognitive systems capable
3GSM World Congress in February 2002, of assisting or replacing soldiers on
Mitsubishi Electric demonstrated a SIM card hazardous duty or civilians responding
featuring voice validation software to toxic spills or disasters.
developed by Domain Dynamics Ltd. The
software provides a ‘biometric template’ that In addition to AIs that focus on novel ways
can recognise a person’s voice to properly of learning, other programmes exist which
identify a user – ‘a necessity when providing can be said to primarily reason. Perhaps the
access to corporate or private databases over most successful example in operation today
the Internet’ (Mokhoff, 2002). With regard is the SmartAirport Operations Centre,
to speech recognition’s natural successor, a logistics programme created by Ascent
natural language processing, such Technology. This AI uses genetic algorithms
technology is, to date, poorly developed and to plan airport timetables by calculating
computers are not yet able to even approach how to optimise complicated scenarios.
the ability of humans to extract meaning Other reasoning programmes are based on
from natural languages (Stottler Henke, heuristic classification – a form of expert
2002). However, due to the many potentially system – and are generally considered the
52
functionality that exceeds the sum of their products, rather than their degree of
parts (The Disappearing Computer, 2003). It intelligence.
is from these ideas that the concept of robot
teams begins to emerge. Robot teams As alluded to above, one of the most
potentially have applications in a wide range commercially valuable frontiers of AI is
of areas. This is because robots working in e-commerce, where technicians are hoping
teams ‘allow for solutions in which to make the online world simpler and more
knowledge, expertise, and motor capability capable at the same time. Robots, too, are
may be distributed in time and space’ (Maes, potentially big business for the hi-tech
1994). Thus, while individual robots may companies prepared to invest in them:
only have limited capacity, robots working investment in robots world-wide increased
together in groups might be able to perform markedly during 2000, with almost 100,000
complex tasks. These include military new units being installed, raising the total
surveillance, mine removal, automated stock of robots to 750,000 at the end of
household tasks, large scale laboratory 2000 (The Economist, 2001).
projects (such as those used in the Human
Genome Project) and assembly. In this way, 3.3.6.1 The US and Japan
most military planners believe that robots In general, the US is more widely regarded for
and remote-controlled sensors represent the its private software development than it is for
future of information collection on the its hardware, for which Japan is most highly
battle-field (Jeremiah, 1995). thought of (Shim, 2002). Indeed, as noted
earlier, the number of US AI-related patents
3.3.6 Corporate funding in existence increased from 100 in 1989, to
While Section 3.3.5 has demonstrated the 1,700 in 1999. Private firms, including large
significant commercial interest in AI, the manufacturers of electronics and computers,
picture for corporate investment in this area as well as major users of IT, hold a vast
is a far less coherent. To date, unlike the field majority of these patents. The top three of
of nanotechnology, no significant overview of these are IBM (297 patents), Hitachi (192)
AI funding seems to exist in the literature. and Motorola (114), although another 17
Having said this, however, the level of companies make an appearance on the list7.
corporate support for AI application Similarly, many of Japan’s major companies
development is, in all likelihood, have plans for AI. According to Shim (2002),
considerable: according to Henry McDonald, the trick for major companies ‘is to time things
Director of the NASA Ames Research right so as to be on the cutting edge of the next
Centre, one-third of computer-science big thing.’ For example, one area in which
funding comes from government and two- Sony – one of the most successful companies
thirds from industry (cited in Krill, 2002). in the history of consumer electronics – has
This is not to say, though, that the interests invested in heavily is the home-robot market,
of the scientific and business worlds through its Entertainment Robot America
necessarily concur; while AI may pose many division. Sony’s latest development in this area
fascinating questions for the former, such concerns Aibo Recognition, a mechanical dog
technology has to be commercially viable in granted with the ability to recognise its owner’s
the latter (Broersma, 2001). For this reason, name, voice and face, as well as automatically
no industry has yet identified a strong motive recharge itself. By infusing Aibo with increased
for developing strong AI and it is unlikely AI, such as voice and face recognition, the
that scientists and business people will get hope is to give Aibo owners the ability to
any closer together in the future. The central interact with a robot at an unprecedented
focus here, then, must be on the utility of level (Spooner, 2002).
54
strong AI research are far more mundane, Brooks (2002), computer vision systems can
having been developed as a result of new do a few things with great skill, but still after
scientific interest in the mechanisms of the 40 years of effort they are not good at the
brain and the way they learn, evolve and things humans and many animals do
develop intelligence from a sense of being effortlessly. Secondly, robots lack the
conscious. To begin with, although dexterity of the human hand, a primary
computers are certainly becoming faster, such ingredient in the types of manufacturing that
achievements do not necessarily correspond have moved to low-cost locations. According
with computers becoming more intelligent. to Brooks, ‘low-cost dextrous manipulation’
For, as described by Jaron Lanier (cited in is essential if progress is to be made. At
Ho, 2002c) as the ‘great shame’ of computer present, however, even high-cost dextrous
science, Moore’s law in hardware manipulation is beyond researchers.
development must be starkly contrasted with Furthermore, such challenges are unlikely to
the fact that computer engineers do not seem be met in the next few years, possibly
to be able to write software much better as requiring 30–40 years before such
computers get more advanced. technologies are refined.
So far, this report has largely focused on 3.4.3 A future for strong AI?
ways in which scientists model part of what In spite of the many fundamental barriers
we know about our capabilities as sentient highlighted above, the fields of AI and
beings, rather than attempting to provide robotics are replete with many wonderfully
true sentience. However, even if the ability to inventive predictions, a domain where reality
programme software advances rapidly within and science fiction often meet. Indeed, it is
the next few decades, it seems likely that the likely that in the next two decades ‘we’ll see
AI laboratories of the day will be incapable more and better capabilities that we tend to
of providing the kind of environment attribute as awareness’ (Hendler, 2000).
necessary for generating anything resembling However, it is unlikely that machines will ever
well-rounded intelligence. This idea stems have human awareness in the philosophical
largely from the work of Rodney Brooks of sense of the term, although they may come
the MIT who has worked hard in recent close in the long term. Rather, we can expect
years to challenge prevailing attitudes to see classical AI going on to produce more
towards AI research 8. Humphrys (1997) and more sophisticated applications in
builds on these ideas by asserting that you restricted domains, such as expert systems,
can’t expect to build a single, isolated AI chess programs and Internet agents. At the
alone in a laboratory and expect to simulate same time, the next 30 years will produce
much intelligence. This is because, unless AIs new types of animal-inspired machines that
are provided with space in which to evolve a are more ‘messy’ and unpredictable than any
rich culture, with repeated social interaction we have seen before – less rationally
with things that are like them, you cannot intelligent but more rounded and whole
really expect to get beyond a certain stage. (Humphrys, 1997).
56
pre-empt and defeat terrorist acts.’ The tools potential for software and robot autonomy.
which the TIA intends to develop to achieve In the short term, some commentators
this rely to a large extent on new AI question whether people will really want to
technologies. These include ‘entity extraction cede control over our affairs to an artificially
from natural language text’ and ‘biologically intelligent piece of software, which might even
inspired algorithms for agent control.’ have its own legal powers. Broersma (2001)
Furthermore, one of the TIA’s 13 subdivisions, believes that, while some autonomy is
the Human Identification at a Distance beneficial, absolute autonomy is frightening.
(HumanID) programme, is releasing contracts For one thing, it is clear that legal systems are
for face, iris and gait recognition. Another of not yet prepared for high autonomy systems,
the subdivisions, FutureMap, will concentrate even in scenarios that are relatively simple
on market-based techniques for avoiding to envisage, such as the possession of personal
surprise and predicting future events information. In the longer-term, however, in
(Hertzberg, 2002). which it is possible to envisage extremely
advanced applications of hard AI, serious
A second programme, called Evidence questions arise concerning military conflict,
Extraction and Link Discovery (EELD), and robot ‘take-overs’ and machine rights.
aims to develop technology for ‘automated Each of these is dealt with in turn below.
discovery, extraction and linking of sparse
evidence contained in large amounts of 3.5.3.1 AI and military conflict
classified and unclassified data sources’ This report shows that the military interest
(Anthes, 2002). In order to achieve this, in AI is significant. However, as pointed out
EELD will have to develop detection above, the difficulties involved in achieving
capabilities to extract relevant data and anything resembling hard AI surely mean that
relationships about people, organisations any such system will be subject to reliability
and activities from huge volumes of data. concerns. This idea is not new; the issue is
picked up by Thompson as early as 1977, who
Apart from the sheer ambitiousness of the sets out his concerns regarding existing and
programmes, TIA and EELD have generated planned uses of computer technology as part
concern mainly in relation to their of nuclear weapons systems. More generally,
implications for infringing individual and it is his belief that no computer system has the
group privacy, and the possibility of such capacity to reliably make decisions of the
information being handled carelessly or even required kind and in the required
leading to malevolence. Indeed, it only takes a circumstances, nor can one ever be constructed.
moment of reflection to consider that nearly This is because the complexity and sensitivity
everyone in modern society has at least one of such systems makes exhaustive
fact about themselves to hide. And yet, in characterisation extremely difficult, and any
spite of these well-founded concerns, both the resulting mistakes cannot be corrected via the
TIA and EELD are already in active usual process of use, failure and modification.
development; in response, Hertzberg (2002) More recently, the controversial US National
recommends that, at a minimum, a temporary Missile Defence programme, which is being
shutdown of the EELD system pending some designed using the latest AI technology,
sort of congressional review and the creation provides a second example. The system is
of safeguards is highly desirable. supposed to dispense ‘kill power’ based on
an ability to recognise incoming missiles in a
3.5.3 AI and robotic autonomy matter of seconds and then decide whether to
Many of the major ethical issues surrounding destroy, intercept of ignore them (Newquist,
AI-related development hinge upon the 1987). However, serious concerns are already
58
arbitrators in decision making.’ In any case, major applications of AI research are mostly
it is not clear in the first place why hidden from view because they are embedded
intelligence should necessarily be regarded in larger software systems. Second, many of
as synonymous with aggression. On the other these applications are morally ambiguous –
hand, cultural reliance, in which humans a grey area of ethics that stands in stark
somehow allow a position of dependency on contrast to Isaac Asimov’s famously clear-cut
AI and robotics to develop, and co-evolution, three laws of robotics 9. Third, presuming
in which human and machine become that a public debate over AI can be initiated,
inextricably intertwined, are regarded as there is little evidence to date that this
more probable. discussion will affect military and
commercial interests. Having said that, there
The strong public reaction to machine take- is evidence of some attempt to flesh out a
over appears, then, not to be well founded. code of professionalism for AI. For example,
However, if it is possible to agree, for in reference to AI and responsibility, Whitby
argument's sake, that humankind will be (1984) writes:
able to create a truly intelligent machine, a
much deeper issue arises: how will a sentient ‘Where an AI system is introduced into any
artificial being be received by humankind human system it shall be the responsibility
and by society? Barry (2001) asks pertinent of the AI professional to ensure that a
questions: ‘Would it be forced to exist like its human or group of humans within the
automaton predecessors who have effectively system shall take moral and/or legal
been our slaves, or would it enjoy the same responsibility for the human consequences
rights as the humans who created it, simply of any malfunction of the AI system.’
because of its intellect?’ This is an enormous
question that touches religion, politics and However, there is little sign in the literature
law, but to date little serious discussion has that suggests these ideas have been followed
been given to the possibility of a new up on.
intelligent species and to the rights an
autonomous sentient might claim. Strong AI, on the other hand, asks much
more fundamental questions as the field
3.6 Discussion necessarily deals with human/machine
The short-term concerns surrounding AI relationships per se. As a consequence, the
and robotics are mainly ethical in nature. kinds of tools that might be necessary to
This is in contrast to nanotechnology, the begin debate over strong AI are not even here
potential dangers of which cover a much yet, so great are the implications. However, it
larger spectrum and one that includes is likely that this technology will not occur in
environmental risk. As shown above, weak our lifetimes; regardless of how often
AI tends to create concern with respect to Professor Warwick is presented as an AI
its role as a tool for human interaction, expert, the fact remains that his opinions are
throwing up issues of responsibility, far removed from the majority view of the
privacy and trust. Applications in this area AI community (Colton, 2001). On the other
are emerging all the time, making 2003 the hand, this report is by no means intended to
right time to begin public debate over these downplay such potentially revolutionary
concerns. This is important for three main developments as ‘mere’ science fiction. For,
reasons. First, there might be a tendency for if the long-term potential of AI was to be
AI technology to creep into out lives largely realised, then it would surely have a
unnoticed. This is because of the well- demonstrable impact in a whole range of
documented AI effect, due to which the industrial and, in particular, service sectors.
This report began by stressing the need to example, the concourse of nanoscience,
provide background information on biotechnology, IT, and cognitive science
nanotechnology and AI. In doing so, it was (‘NBIC’) was discussed during a December
hoped that the prospects of these emerging 2001 NSF workshop. NBIC, it was agreed
technologies to affect quality of life in the ‘could achieve a tremendous improvement in
coming decades could be realistically human abilities, societal outcomes, the
assessed. One consequence of providing nation’s productivity and the quality of life’
such an overview is that there can be no (Roco and Bainbridge, 2003). In some ways,
decisive conclusions as such; the industries the above conclusion is hardly surprising
characterised here are too dynamic and given the ambitious and broad scope of the
uncertain to generate any real sense of technologies discussed in this report. As
resolution. However, it is possible to pointed out above, ‘convergence’ largely
highlight a number of important differences arises from the wide availability of
and similarities between nanotechnology techniques and tools on offer today – the real
and AI which go some way to shedding innovation stems from the process of
more light on their character. bringing individuals from traditionally
separate disciplines together.
Perhaps the greatest contrast between the
two industries concerns public interest. Most importantly for convergence here,
Indeed, as this report has demonstrated, it is possible that developments in
nanotechnology is widely regarded as a nanotechnology could lead to advances
‘new’ and exciting branch of science and in AI through improvements in computer
technology. This belief has contributed to miniaturisation, performance, or architecture
the massive period of growth that this high- (but see Section 3.4.2 on barriers to strong
profile and wide-ranging field is currently AI), or through the sensor interface. In
enjoying. AI, on the other hand, is viewed by addition, it seems fair to assume that any
many as an highly specialised and unproven futuristic nanobots would have to be imbued
discipline. One reason for this concerns the with a reasonable degree of AI. A second,
gross over-optimism that characterised the more contentious similarity concerns
industry in the 1960s and 1980s. Another reinvention. As demonstrated in this report,
reason reflects the AI community’s seemingly the ‘rediscovery’ of AI has been a virtual
insurmountable difficulty in publicising its necessity for the survival of the industry;
own achievements without whipping up for nanotechnology the phenomena is less
general anxiety over machine superiority. obvious but is arguably there all the same.
The upshot of all this has been the field’s That is, as a natural extension of the
struggle to attract funding in the past and micromechanical and MEMS research that
it is likely that this trend will continue for begun in the 1960s, nanoscience is hardly
sometime into the foreseeable future. ‘new’ as such; rather, ‘nano’ can be viewed
as a useful tag with which to boost funding.
Revealing similarities also exist between Just what the consequences of this strategy
nanotechnology and AI. There has been will be, it is hard to tell. Ironically, AI
much talk recently regarding the convergence provides an excellent example of a promising
of traditionally separate scientific fields, in scientific discipline that has often resulted in
particular the blurring of the boundaries disappointment. Whether the same happens
between the physical sciences and life to nanotechnology remains to be seen.
sciences – perhaps even the first step towards
the long sought after unification of physics, The second consequence of providing an
chemistry and biology (Howard, 2002). For overview is that certain elements of
60
nanotechnology and AI development are by industry, transport and the domestic
bound to be overlooked. First, the difficulties sector. The way in which these more
of drawing out accurate statistics for fundamental changes might impact on the
corporate R&D have already been alluded environment would have to form the basis
to earlier. Second, there are wide ranging of a much larger technology assessment, in
applications across the economy for sensors which long-term structural changes to global
that can support industrial processes and be industry and commerce were considered.
incorporated into new or existing products
(Miles and Jarvis, 2001). The application of Finally, it is easy to overlook the lessons that
nanotechnology to this area should allow for attitudes towards technological development
improvements in functionality and much teach us about human nature. This report has
decreased size. Third, a more in-depth largely relied upon the technique of looking
analysis of environmental concerns is ahead, identifying technological possibilities,
warranted. This is because public and assessing the likelihood of successfully
acceptability of such risk is likely to vary moving towards their realisation. Significantly,
considerably in relation to the application this process mirrors that of technological
being considered. For example, the innovators, a kind of thinking that often
application of nanotechnology to translates into the belief that technological
computerisation is less likely to cause development is autonomous – the ultimate
concern than those practices which might self-fulfilling prophecy. To some extent we are
lead to the release of nanoparticles into the already on this road. Most technologies
environment, such as the disposal of nano- covered in this report are within the bounds of
based composites. Fourth, it is possible to current scientific possibility and it is just a
conceive of a number of environmental matter of time, effort and expenditure before
goods that may arise. For example, the they are realised. However, the contrasting
potential for gains in energy generation and fortunes of the nanotechnology and AI
efficiency have already pointed out above industries remind us that much of this
(Section 2.3.4.), and it is conceivable that progress hinges on public approval.
dramatic improvements in environmental Ultimately, a 21st-Century acceptance model
sensing and modelling could also be calls for technological innovations to be
achieved. However, any pervasive diffusion received on a voluntary basis where the
of nano- and AI-based technologies in the perceptible usefulness of new technology
coming decades is bound to have a products are balanced against associated risks
significant effect on the demand for resources that are shown to be manageable.
1
Grove-White, R., Macnaghton, P. and Wynne B. (2000). Wising Up: The Public
and New Technologies, Lancaster, UK: IEPPP, Lancaster University.
2
It is worth bearing in mind when consulting this type of information that, given
the difficulty of even agreeing on what constitutes nanotechnology, many of the
numbers presented below should be treated with caution (Roman, 2002).
3
For a detailed breakdown, see the NNI’s own report at:
http://www.nano.gov/2003budget.html
4
For a detailed description of the history of AI, see the University of Edinburgh’s
Division of Informatics Website at:
http://www.dai.ed.ac.uk/AI_at_Edinburgh_perspective.html
5
For more information, see Ruthfield, 1995.
6
See Honda’s Website for more information at: http://world.honda.com/robot/
7
For details, see National Research Council, 1999.
8
For a full description of this paradigm shift, see Humphrys, 1997.
9
The Three Laws of Robotics are:
a. A robot may not injure a human being, or, through inaction, allow a
human being to come to harm;
b. A robot must obey the orders given to it by human beings except where
such orders would conflict with the First Law;
c. A robot must protect its own existence as long as such protection does not
conflict with the First or Second Law.
62
References A–C
Albery, N. (1999). Robot Terror. .BBC (2002). Computers Reach Chapman, G. (2000). Quoted in:
New Scientist; 161 (2171): 49. One Million Mark. BBC News; 1 Matthews, J. (2000b). Military
[letter] Jul 2002 [online]. Applications of AI. Generation5
http://news.bbc.co.uk/1/hi/sci/tech/ Essays [online].
Aleksander, I. (2002). The Next 2077986.stm http://www.generation5.org/app_
Big Thing. BBCi & The Open military.shtml
University [online]. Boeing (2002). Unmanned
http://www.open2.net/nextbigthing Combat Air Vehicle (X-45). Chaudhari, P. (2001). Future
/ai/hear_the_arguments/argument3 Boeing [online]. Implications of Nanoscale Science
.htm http://www.boeing.com/phantom/u and Technology: Wired Human,
cav.html Quantum Legos, and an Ocean of
Anderson, K. (2001). Predicting Information. In: M.C. Roco, W.S.
AI’s Future. BBC News; 21 Sep Broersma, M. (2001). AI Gets Bainbridge (editors). Societal
2001 [online]. Down to Business. ZDNet UK; Implications of Nanoscience and
http://news.bbc.co.uk/1/hi/in_dept News: 23 Jan 2001 [online]. Nanotechnology. Arlington, VA,
h/sci_tech/2001/artificial_intelligen http://news.zdnet.co.uk/story/0,,s2 USA: National Science
ce/1555742.stm 083916,00.html Foundation; pp75–9.
http://www.wtec.org/loyola/nano/
Antón, P.S. Silberglitt, R. and Brooks, R. (2001). Interviewed in: NSET.Societal.Implications/nanosi.
Schneider, J. (2001). The Global Anderson, K. Predicting AI’s pdf
Technology Revolution: future. BBC News; 21 Sep 2001
Bio/Nano/Materials Trends and [online]. Colton, S. (2001). Cyborg Off His
Their Synergies with Information http://news.bbc.co.uk/1/hi/in_dept Christmas Tree. Times Higher
Technology by 2015. Santa h/sci_tech/2001/artificial_intelligen Education Supplement; 21 Dec
Monica, CA: RAND. Document ce/1555742.stm 2001: p14.
Number: MR-1307- http://www.aisb.org.uk/articles/co
NIC.http://www.rand.org/publicati Brooks, R. (2002). Interviewed in: mputology.html
ons/MR/MR1307/ Jusko, J. The Robot Evolution.
Industry Week; 1 Nov 2002 Colvin, V. (2002). Responsible
Anthes, G.H. (2002). Global [online]. Nanotechnology: Looking Beyond
Surveillance: The Government’s http://www.industryweek.com/Cur the Good News. EurekAlert!;
Plan. Computerworld; Security: rentArticles/Asp/articles.asp?Articl Nanotechnology In Context: Nov
Privacy, 25 Nov 2002 [online]. eId=1356 2002 [online].
http://www.computerworld.com/se http://www.eurekalert.org/context.
curitytopics/security/privacy/story/ Buchanan, B. and Uthurusamy, S. php?context=nano&show=essays
0,10801,76117,00.html (1999). The Innovative &essaydate=1102
Applications of Artificial
Augarten, S. (1986). Can Intelligence Conference. AI Compano, R. editor (2001).
Computers be Failsafe? PC Magazine; 20 (1): 11–12. Technology Roadmap for
Magazine; 5 (1): 97. Nanoelectronics. Luxembourg:
Cameron, D. (2002). Rat-Brained Office for Official Publications of
Barry, R. (2001). Sentience: The Robot. Technology Review; 18 the European Communities.
Next Moral Dilemma. ZDNet Dec 2002 [online].
UK; News: 24 Jan 2001 [online]. http://www.technologyreview.com/
http://news.zdnet.co.uk/story/0,,s2 articles/print_version/wo_cameron
083942,00.html 121802.asp
Compano, R. and Hullman, A. Dertouzos, M.L. (1999). The Economist, The (2002). Trouble in
(2002). Forecasting the Future of Computing. Scientific Nanoland. The Economist.com; 5
Development of Nanotechnology American; 281 (2): 52–5. Dec 2002 [online].
with the Help of Science and http://www.sciam.com/issue.cfm?is http://www.economist.com/science/
Technology Indicators. suedate=Aug-99 displaystory.cfm?story_id=1477445
Nanotechnology; 13 (3): 243–7.
Disappearing Computer, The Engineering and Physical Sciences
Copeland, B.J. (2000). What is (2003). Mission Statement. The Research Council (2003). EPSRC
Artificial Intelligence? Disappearing Computer [online]. Web Site [online].
AlanTuring.net; Reference http://www.disappearing- http://www.epsrc.ac.uk
Articles: May 2000 [online]. computer.net/
http://www.alanturing.net/turing_a ETC Group (2002a). No Small
rchive/pages/Reference%20Articles Doering, R. (2001). Societal Matter! Nanotech Particles
/what_is_AI/What%20is%20AI02. Implications of Scaling to Penetrate Living Cells and
html Nanoelectronics. In: M.C. Roco, Accumulate in Animal Organs.
W.S.Bainbridge (editors). Societal ETC Group Communique; (76)
Cowen, R. (2002). Ribbon to the Implications of Nanoscience and May/June 2002 [online].
Stars. Science News; 162 (14): Nanotechnology. Arlington, VA, http://www.etcgroup.org/documen
218. http://www.sciencenews.org/ USA: National Science ts/Comm_NanoMat_July02.pdf
20021005/bob9.asp Foundation; pp68–75.
http://www.wtec.org/loyola/nano/ ETC Group (2002b). Nanotech
Dalesio, E.P. (2002). ‘Text mining’ NSET.Societal.Implications/nanosi. Takes a Giant Step Down! ETC
Software Business Grows. Sun- pdf Group News Release; 6 Mar 2002
Sentinel; 4 Mar 2002 [online]. [online].
http://www.sun- Doyle, J. and Dean, T. (1996). http://www.etcgroup.org/documen
sentinel.com/business/local/sfl- Strategic Directions in Artificial ts/nr2002march6.pdf
0304textmining.story?coll=sfla- Intelligence. ACM Computing
business-headlines Surveys; 28 (4): 653–70. ETC Group (2002c). Patenting
http://medg.lcs.mit.edu/ftp/doyle/s Elements of Nature. ETC Group
Department of Defense (2002). dai96.html Geno-type; 25 Mar 2002 [online].
Nanotechnology Research Awards http://www.etcgroup.org/documen
Announced. DefenseLINK; News: Dreyfus, H. (1992). What ts/nanopatentsgeno.rtf.pdf
23 Feb 2001 [online]. Computers Still Can’t Do: A
http://www.defenselink.mil/news/F Critique of Artificial Reason. European Environment Agency
eb2001/b02232001_bt079- Cambridge, MA, USA: (2003). EEA Multilingual
01.html Massachusetts Institute of Environmental Glossary.
Technology. European Environment Agency
Department of Trade and Industry [online].
(2002). New Dimensions for Economist, The (2001). Robots. http://glossary.eea.eu.int/EEAGloss
Manufacturing: UK Strategy for The Economist.com; 29 Nov 2001 ary/P/precautionary_approach
Nanotechnology. Report of the [online].
UK Advisory Group on http://www.economist.com/printed Fishbine, G. (2002). Investor’s
Nanotechnology Applications; ition/displayStory.cfm?Story_ID=8 Guide to Nanotech and
June 2002. UK: Department of 87290 Micromachines. Chichester, UK:
Trade and Industry. John Wiley.
http://www.dti.gov.uk/innovation/
nanotechnologyreport.pdf
64
F–H
FOLDOC (2003). Free On-Line Gorman, J. (2002). Taming High- Hay, J.N. and Shaw, S.J. (2000). A
Dictionary of Computing Tech Particles: Cautious Steps into Review of Nanocomposites 2000.
[Online]. the Nanotech Future. Science Proceedings of the
http://wombat.doc.ic.ac.uk/foldoc/ News; 161 (13): 200. Nanocomposites 2000
http://www.sciencenews.org/20020 Conference; 6–7 Nov 2000,
Forrest, D. (1989). Regulating 330/bob8.asp Brussels, Belgium. Farnborough,
Nanotechnology Development. UK: Defence Evaluation and
Foresight Institute; 23 Mar 1989 Graham-Rowe, D. (2002). “Robo- Research Agency.
[online]. rat” Controlled by Brain http://www.nano.org.uk/nanocom
http://www.foresight.org/NanoRev Electrodes. New Scientist; 417: 37. posites_review.pdf
/Forrest1989.html http://www.wireheading.com/robo
rats/ratbot.html Helsel, S. (2002). Are Corporate
Freitas, R.A. (2000). Some Limits and Government Funding
to Global Ecophagy by Biovorous Grosz, B. and Davis, R. editors Overlooked?
Nanoreplicators, with Public (1994). A Report to ARPA on NanoelectronicsPlanet; 26 Jul
Policy Recommendations. Twenty-First Century Intelligent 2002. [online].
Foresight Institute; Mar 2000 Systems. Menlo Park, CA, USA: http://www.nanoelectronicsplanet.
[online]. American Association for com/features/article/0,4028,6571_
http://www.foresight.org/NanoRev Artificial Intelligence. 1433691,00.html
/Ecophagy.html http://www.aaai.org/Resources/Pol
icy/arpa-report.html Herrera, S. (2002). No small
Fried, J. (2002). Japan Sees matter. Red Herring; 16 Dec 2002
Nanotech as Key to Rebuilding its Grupp, H. editor (1993). [online].
Economy. Small Times; 7 Jan Technologie am Beginn des 21. http://www.redherring.com/insider
2002 [online]. Jahrhunderts. Heidelberg: Physica /2002/12/10trends-
http://www.smalltimes.com/docum Verlag. Cited in: Compano, R. and nanotech121602.html
ent_display.cfm?document_id=284 Hullman, A. (2002).
3 Hertzberg, H. (2002). Too Much
Gsponer, A. (2002). From the Lab Information. The New Yorker; 9
Glinos, K. (1999). to the Battlefield? Nanotechnology Dec 2002. [online]
Nanotechnology: Blurring the and Fourth-Generation Nuclear http://www.newyorker.com/talk/co
Boundaries at the Atomic Scale. Weapons. Disarmament ntent/?021209ta_talk_hertzberg
RTD info; (21) Feb 1999 [Online]. Diplomacy; (67) Oct–Nov 2002
http://www.europa.eu.int/comm/re [online]. Hendler, J. (2000). A Chat About
search/rtdinf21/en/dossier1.html http://www.acronym.org.uk/dd/dd the Future of Artificial
67/67op1.htm Intelligence. CNN; 1 Jan 2000
Goodwins, R. (2001). The [online].
Machine that Wanted to be a Harper, T. (2002). The http://www.cnn.com/COMMUNIT
Mind. ZDNet UK; News: 23 Jan Nanotechnology Arms Race: Why Y/transcripts/1999/12/hendler/
2001 [online]. Nobody Wants to be Left Behind.
http://news.zdnet.co.uk/story/0,,s2 nanotechweb.org; 14 Nov 2002 Ho, M.W. (2002a).
083911,00.html [online]. Nanotechnology: A Hard Pill to
http://nanotechweb.org/articles/col Swallow. Science in Society; (16)
umn/1/11/1/1 Autumn 2002 [online].
http://www.i-
sis.org.uk/nanotechnology.php
Ho, M.W. (2002b). The Brave Jeremiah, D.E. (1995). Leo, A. (2001). Get Ready for
New World Quartet. Science in Nanotechnology and Global Your Nano Future. Technology
Society; (16) Autumn 2002 Security. Proceedings of the Review; May 2001 [online].
[online]. http://www.i- Fourth Foresight Conference on http://www.technologyreview.com/
sis.org.uk/bravenewworld.php Molecular Nanotechnology; articles/wo_leo050401.asp
9–11 Nov 1995, Palo Alto, CA,
Ho, M. (2002c). Will Computers USA. [online]. Lighthill, J. (1973). Artificial
Become Super-Human? Science in http://www.zyvex.com/nanotech/n Intelligence: A General Survey. In:
Society; (16) Autumn 2002 ano4/jeremiahPaper.html Artificial Intelligence: A Paper
[online]. http://www.i- Symposium. London, UK: Science
sis.org.uk/computersvshumans.php Joseph, S. (2001). AI. J@pan Inc Research Council.
Magazine; Nov 2001 [online].
Holister, P. (2002). Nanotech: The http://www.japaninc.com/article.p Machan, D. (2002). A Few Good
Tiny Revolution. CMP Cientifica; hp?articleID=514 Toys. Forbes Magazine; (12)
July 2002 [online]. 9 Dec 2002 [online].
http://www.cientifica.info/html/doc Khan, J. (2002). It’s Alive! Wired; http://www.forbes.com/forbes/200
s/NOR_White_Paper.pdf 10 (3) [online]. 2/1209/120.html
http://www.wired.com/wired/archi
Howard, S. (2002). ve/10.03/everywhere.html Maes, P. (1994). Agents That
Nanotechnology and Mass Reduce Work and Information
Destruction: The Need for an Kirsner, S. (2002). Getting Smart Overload. Communications of the
Inner Space Treaty. Disarmament About Predictive Intelligence. ACM; 37 (7): 31–40.
Diplomacy; (65) Jul–Aug 2002 Boston Globe; 30 Dec 2002 http://pattie.www.media.mit.edu/p
[online]. [online]. eople/pattie/CACM-94/CACM-
http://www.acronym.org.uk/dd/dd http://www.boston.com/globe 94.p1.html
65/65op1.htm
Krill, P. (2002). CTO Forum: Will Malcolm, C. (2001). Robots
Hsiung, S. (2002) An Introduction Artificial Intelligence Surpass the Won’t Rule. School of Artificial
to Artificial Intelligence. Human Variety? InfoWorld; 11 Intelligence, Division of
Generation5 [online]. Apr 2002 [online]. Informatics, Edinburgh
http://www.generation5.org/aiintro http://www.infoworld.com/articles/ University; 3 Aug 2001 [online].
.shtml hn/xml/02/04/11/020411hnartifici http://www.dai.ed.ac.uk/homes/ca
al.xml m/Robots_Wont_Rule.shtml
Humphrys, M. (1997). AI is
Possible…But it Won’t Happen: LaVan, D.A. and Langer, R. Martel, S. (2000). Introduction to
The Future of Artificial (2001). Implications of the NanoWalker: A Miniature
Intelligence. Proceedings of the Nanotechnology in the Autonomous Robot Capable of
Next Generation Symposium; Aug Pharmaceutics and Medical Fields. Various Tasks at the Molecular
1997, Jesus College, Cambridge, In: M.C. Roco and W.S. and Atomic Scales. Bio-
UK. http://www.compapp.dcu.ie/ Bainbridge (editors). Societal Instrumentation Laboratory;
~humphrys/newsci.html Implications of Nanoscience and Projects [online].
Nanotechnology. Arlington, VA, http://bioinstrumentation.mit.edu
IT Strategic Headquarters (2001). USA: National Science
E-Japan Strategy. Policy Initiative, Foundation; pp79–83.
22 Jan 2001 Tokyo, Japan: IT http://www.wtec.org/loyola/nano/
Policy Office. NSET.Societal.Implications/nanosi.
http://www.kantei.go.jp/foreign/it/ pdf
network/0122full_e.html
66
M–P
Matthews, J. (1999). Philosophical Miles, I. and Jarvis, D. (2001). National Institute of Informatics
Arguments For and Against AI. Nanotechnology – A Scenario for (2002). National Institute of
Generation5; 13 Dec 1999 Success in 2006. Teddington, UK: Informatics 2002. Tokyo, Japan:
[online]. HMSO. National Physical National Institute of Informatics.
http://www.generation5.org/ai_phi Laboratory Report Number: http://www.nii.ac.jp/publications/n
l.shtml CBTLM 16. ii-yoran/yoran2002-e.pdf
http://libsvr.npl.co.uk/npl_web/pdf/
Matthews, J. (2000a). cbtlm16.pdf National Research Council
Introduction to Neural Networks. (1999). Funding a Revolution:
Generation5; 20 Mar 2000 Miller, D. (2001). Five Government Support for
[online]. Technologies You Need to Know. Computing Research. Washington,
http://www.generation5.org/nnintr The Industry Standard; 21 May DC, USA: National Academy
o.shtml 2001 [online]. Press.
http://www.thestandard.com/articl http://www.nap.edu/readingroom/
Matthews, J. (2000b). Military e/0,1902,24308,00.html books/far/notice.html
Applications of AI. Generation5;
8 May 2000 [online]. Miller, P. (2002). Big, Big Plans for Nelson, M. and Shipbaugh, C.
http://www.generation5.org/app_ Very Small Things. Green Futures; (1995). The Potential of
military.shtml Jul/Aug 2002: 50–1. Nanotechnology for Molecular
http://www.paulmiller.org/Nanotec Manufacturing. Santa Monica,
McCarthy, J. (2003). What is hnology.pdf CA, USA: RAND. Document
Artificial Intelligence? Computer Number: MR-615-RC.
Science Department, Stanford Tam, A. (2001). Japan’s Nanotech http://www.rand.org/publications/
University; 29 Mar 2003 [online]. Players: Thinking Big. J@pan Inc MR/MR615/
http://www- Magazine; May 2001 [online].
formal.stanford.edu/jmc/whatisai/ http://www.japaninc.net/mag/com Newquist, H. (1987). Star Wars:
whatisai.html p/2001/05/print/may01p_investor_ When Imperfect Man Strives to
nanotech.html Make Perfect Machines.
McCullagh, D. (2002). Report Computerworld; 30 Nov 1987:
Calls for Nanotech Laissez-Faire. Mokhoff, N. (2002). Speech 17. http://www-
CNET News.com; 21 Nov 2002 Technology Looses its Kooky cse.stanford.edu/classes/cs201/Proj
[online]. Luster. EETimes; 1 Nov 2002 ects/autonomous-
http://www.news.com.com/2100- [online]. weapons/articles/starwars.txt
1023-966766.html http://www.eetimes.com/story/OE
G20021031S0046 New Scientist (2002). US Army
Melymuka, K. (2002). Good Seeks Nanotech Suits. New
Morning, Dave... Computerworld; nanotechweb.org (2002). Scientist.com; News Service, 14
11 Nov 2002 [online]. Nanocomposites Set to Wrap up Mar 2002 [online].
http://www.computerworld.com/so the Packaging Market. http://www.newscientist.com/news
ftwaretopics/software/appdev/story nanotechweb.org; 23 Aug 2002 /news.jsp?id=ns99992043
/0,10801,75728,00.html [online].
http://nanotechweb.org/articles/ne Pace, S. (1989). Military
ws/1/8/18/1 Implications of Nanotechnology.
Foresight Update 6; 1 Aug 1989
[online].
http://www.foresight.org/Updates/
Update06/Update06.2.html
Pergamit, G. and Peterson, C. Roco, M.C., and Bainbridge, W.S., Schneider, F.B. and Rodd, M.,
(1993). Nanotechnology. In: editors (2002). Converging editors (2001). International
Ringo, T. editor. On The Cutting Technologies for Improving Review of UK Research in
Edge of Technology. Indianapolis, Human Performance. Computer Science. Swindon, UK:
IN, USA: Sams Publishing. Nanotechnology, Biotechnology, Engineering and Physical Sciences
Information Technology and Research Council.
Philipson, G. (2001). Human and Cognitive Science. Arlington, VA, http://www.iee.org/policy/csreport/
Computer Merger a Chip of the USA: National Science cs_report.pdf
Old Block. The Age; IT News, 18 Foundation.
Sep 2001, p1. For a summary see: http://wtec.org/ConvergingTechnol Schulz, W.G. (2002).
http://www.gyre.org/news/related/ ogies/Report/NBIC_pre_publicatio Nanotechnology Under the Scope.
Metacomputing/Biological+Comp n.pdf Chemical and Engineering News;
uting 4 Dec 2002 [online].
Roman, C. (2002). It’s Ours to http://pubs.acs.org/cen/today/dec4.
RAND (2002). Revolutionising Lose: An Analysis of EU html
Logistics Support by “Thinking Nanotechnology Funding and the
Small”. RAND; National Security Sixth Framework Programme. Science Blog (2002). New Centre
[online]. Brussels, Belgium: European to Transfer Nano Innovations to
http://www.rand.org/natsec_area/p NanoBusiness Association. Defence Industry. Science Blog;
roducts/thinkingsmall.html http://www.nanoeurope.org/docs/E 10 Dec 2002 [online].
uropean%20Nanotech%20Fundin http://www.scienceblog.com/comm
Reynolds, G.H. (2002). Forward g.pdf unity/article553.html
to the Future: Nanotechnology
and Regulatory Policy. San Roy, R. (2002). Giga Science and Shim, R. (2002). Sony’s Ando: PCs
Francisco, CA, USA: Pacific Society. Materials Today; to Function Like a Brain. ZDNet;
Research Institute. 5 (12): 72. News: 5 Dec 2002 [online].
http://www.pacificresearch.org/pu http://www.materialstoday.com/pd http://zdnet.com.com/2100-1105-
b/sab/techno/forward_to_nanotech fs_5_12/opinion.pdf 976269.html
.pdf
Ruthfield, S. (1995). The Internet’s Smalley, R.E. (2001). Of
Roco, M.C., and Bainbridge, W.S., History and Development. AMC Chemistry, Love and Nanobots.
editors (2001). Societal Crossroads; 2 (1) [online]. Scientific American; 285: 76–7.
Implications of Nanoscience and http://www.acm.org/crossroads/xr http://www.ruf.rice.edu/~smalleyg/
Nanotechnology. Arlington, VA, ds2-1/inet-history.html rick%27s%20publications/SA285-
USA: National Science 76.pdf
Foundation. Saxl, O. (2000). Opportunities for
http://www.wtec.org/loyola/nano/ Industry in the Application of Smith, R.H. (1996). Molecular
NSET.Societal.Implications/nanosi. Nanotechnology. London, UK: Nanotechnology: Research Funding
pdf Office of Science and Technology. Sources. Nanotechnology Magazine;
[A report for The Institute of 2 (6).
Nanotechnology, April 2000].
http://www.nano.org.uk/contents.h
tm
68
S–Z
Smith, R.H. (2001). Social, Trevelyan, J. (1999). White House Fact Sheet
Ethical, and Legal Redefining Robotics for the (2000). National
Implications of New Millennium. Nanotechnology Initiative:
Nanotechnology. In: Roco, International Journal of Leading to the Next
M.C., and Bainbridge, W.S., Robotics Research; 18 (12): Industrial Revolution.
editors. Societal Implications 1211–23. Washington, DC, USA: The
of Nanoscience and White House.
Nanotechnology. Arlington, Turing, A. (1950). http://clinton4.nara.gov/texto
VA, USA: National Science Computing Machinery and nly/WH/New/html/20000121
Foundation. pp203–11. Intelligence. Mind; 9: _4.html
http://www.wtec.org/loyola/n 433–60.
ano/NSET.Societal.Implicatio http://cogprints.ecs.soton.ac. Wolbring, G. (2002). The
ns/nanosi.pdf uk/archive/00000499/00/turi Silenced Targets. In: Institute
ng.html of Science in Society Special
Spooner, J.G. (2002). Sony Series – Inside Human
Teaches Abio New Tricks. Weld, D.S., editor. (1995). Genetics and Genomics; Part
ZDNet; News: 10 Oct 2002 The Role of Intelligent II: 28 Jan 2002 [online].
[online]. Systems in the National http://www.i-
http://zdnet.com.com/2100- Information Infrastructure. sis.org.uk/SilencedTargets.php
1106-961536.html American Association for
Artificial Intelligence Zyvex (2002). Zyvex Web
Stottler Henke (2002). [online]. Site [online].
Glossary of AI Terms. http://www.aaai.org/Resourc http://www.zyvex.com/Produ
Stottler Henke Associates, es/Policy/nii.html cts/tools.html
Inc. [online].
http://www.shai.com/ai_gene Whitby, B. (1984) A Code of
ral/glossary.htm Professionalism for AI. AISB
Quarterly; (64) [online].
Thibodeau, P. (2002). Senate http://www.aisb.org.uk/articl
Scrutinizes US Nanotech es/professionalism.html
Investment. Computer
World; 18 Sep 2002 [online]. Whitby, B. and Oliver, K.
http://computerworld.com/m (2001). How to Avoid a
anagementtopics/managemen Robot Takeover: Political
t/itspending/story/0,10801,74 and Ethical Choices in the
341,00.html Design and Introduction of
Intelligent Artifacts. AISB
Thompson, H. (1977). There Quarterly; (104) [online].
Will Always Be Another http://www.aisb.org.uk/articl
Moonrise: Computer es/avoid_takeover.html
Technology and Nuclear
Weapons. AISB Quarterly;
(53/54) [online].
http://www.aisb.org.uk/articl
es/moonrise.html