Professional Documents
Culture Documents
10-12/30-4/2019
A barren landscape, all that we’ve come to know, and love is gone now, man versus
machines for survival. Ruins may be all that is left for the humans, as the war has left nothing
else for them. The only human hope is through one man and his resistance, John Connor. What
I’ve just explained to you, is the opening of Terminator 2: Judgement Day, which shows the
audience a take on AI dominance. In this case, an AI had become sentient and ruled that humans
must be destroyed if they seek to shut it down. This movie provides a common perspective
towards AI today, that was instilled effectively as a result of the series. Looking at the second
movie alone, it made about $524 million (MOJO) at the box office, which is about $975 million
after inflation (WESTEGG). This preconception of an apocalyptic future is only but the
nightmare situation, and it didn’t start here, one of the earliest forms of this idea was addressed
Before talking about Darwin Among the Machines, it’s only appropriate to define what
AI is and the “father” of AI. According to Encyclopedia Britannica, Artificial Intelligence is,
associated with intelligent beings.” The modern-day definition is derived from Alan Turing, who
including the artificial neural network system. His contribution to the field of AI was so
staggering that they hold a challenge for those interested in AI. The challenge is called the
“Turing test.” The “Turing test” is a challenge that tasks humans to create an AI that is to be
indiscernible from humans in intelligence. This test is carried out through three "players,” the
first being the AI, the second being a human interrogator, and the third being a human player.
The test is carried out by having the interrogator ask questions to both the computer and the
human foil. The computer and the human answer these questions on paper and give them back to
the interrogator. Once the interrogator is done examining the answers, they should not be able to
tell which is the human and which is the machine. To this day, no AI has convincingly passed
this test while being moderated. There being no official winner of the “Turing test,” can lead to
believe that no sentient AI has been built. But just because the AI isn’t sentient, doesn’t define its
Take the AI chess master “Deep Blue,” for instance. “Deep Blue,” was an AI specifically
made for chess by IBM and could take on the world champion of chess at the time, Garry
Kasparov. “Deep Blue,” wasn’t made for any other task than chess, though, it wasn’t sentient,
being that it couldn’t think like a human. “Deep Blue” only knew how to play chess. It only
knew the algorithms and scenarios that were programmed to respond to any move Kasparov
made. (Copeland) “Deep Blue” is a prime example of how AI can be intelligent but not sentient.
Deep Blue wouldn’t have been able to pass a “Turing test,” due to its inability to think like a
human. It would only be able to analyze chess moves and calculate the best route to a victory
humans allow it to be so? If so, what will be AI’s relationship with humans? Will it lead to
human destruction, or will it lead to vast improvements in human civilization?” No one knows
“Darwin Among the Machines,” leads to reason that this argument has existed before but
is one of the earliest written forms of the idea of “machine overlords.” Butler makes the point
that mankind has evolved to this point and has developed to be above all the other species and
tries to answer the question, “If humans are superior now, who will eventually rule over
humans?” The conclusion that Butler comes to is that humans will eventually create their own
rulers in the form of machines. Butler goes further with this idea though and relates it to how
humans have been merciful to animals like horses, dogs, and cats because they do not endanger
us and serve some function that is useful to humans. This idea of co-inhabitance is what Butler
insists, where these “mechanical rulers” will be merciful because they will need to depend on us
view but flawed if not clarified. He seems to take the approach that these robots will be alive,
just like an animal and therefore will need us to sustain them. If the apocalyptic scenario were to
be played out, then he would be wrong, robots would try and establish their own way and
wouldn’t need the help of humans. On the other hand, if humans establish robots with safety
standards, there would be that sort of mutualistic relationship that Butler points out. I align
myself with the latter half, the key to a prosperous future with AI starts with cautious humans.
Whether it be humanity or just the job market, humans must prepare themselves to be successful,
AI in the job market were, she was convinced that robots would end up replacing us. Her
thinking went above the job market, but she thought that robots would end up replacing us in our
entirety. I aligned myself with that perspective as well as I was a huge science fiction fan. It took
a lot of convincing from multiple sources to assure me that robots shouldn’t evolve to the point
where they do not need humans. According to the Future of Life Institute, AI should
theoretically only be dangerous if it is developed with malintent or given instructions that could
be misunderstood. It’s important to address that this is only a theory, the theoretical and actual,
can be quite different. Despite this point, though the Future of Life Institute still makes the point
that, if humans make AI safe, there is no need to worry, especially now, where the goal of
creating sentient AI is far off into the future. Applied to the manufacturing industry, even if new
technology is introduced, machines that can repair and think for themselves are still a couple of
decades from now. This means that even if new forms of automation are introduced now, there
will still be a need for human assistance with those jobs to help maintain the machines. Then
when they do become sentient, there would still be people needed to monitor and help the AI as
well, creating new jobs to replace the old ones. This phenomenon of new jobs replacing the jobs
lost from automation forms into a sort of cycle that has happened throughout history. The most
popular example of an era, where new jobs replaced old ones phased out by automation is the
Industrial Revolution.
A study published in 2019 by Mark Muro, Robert Maxim, and Jacob Whiton, examined
the cycle of manufacturing jobs impacted by automation throughout history. Muro and his
colleagues are research scientists at the Brookings Institution, which is a nonprofit public policy
organization that researches and looks to solve problems that may face society. (Brookings)
Examining eras within the United States like the Industrial Revolution, the “IT era,” and the new
“AI era.” Muro and his colleagues studied the history of these eras and used it to predict the
potential impact of the “AI era.” These examinations lead to believe that if humans are willing to
work, automation will never fully force humans out of the workplace. Muro and his team believe
that automation has six basic tendencies: “Automation substitutes for labor, automation also
complements labor, machines substitute for tasks, automation can increase demand, capital and
labor augmentation spurs innovation, and technological possibility is not the same as
This preconception of AI, possibly eliminating all jobs in a certain field, such as
manufacturing, is not say impossible, but unlikely due to these six tendencies combined with the
limiting of AI proposed by the Future of Life Institute. To help understand that, it would be best
to go through and explain what these tendencies are as described by Muro and his colleagues’
work at Brookings.
The first tendency, “Automation substitutes for labor,” mostly sets the theme for the rest
of the tendencies. This tendency brings up the point of how machines are overall better at doing
tasks at both faster speed and with greater precision than a human, all for a lower cost. The
problem that they claim to be with using machines is that they can usually only handle one task
at this moment and would not be able to take the place of a human who can do multiple tasks.
There is also a second factor that Muro and his team bring up, which is about supply and
demand. Muro and his team also bring up that with automation, it can cause a decline in wages,
which would allow for a more competitive environment, as well as producing more leads to other
adjustments in pricing.
The second tendency, “Machines substitute for tasks, not jobs,” relates to the first, talking
about how machines cannot do everything that a human can do at this point. Muro and his crew
define a job as “a collection of tasks,” where many things must be done before a job is complete.
They go as far as to say that some tasks are not suited towards machines but instead can be done
better by a human. This leads the group of researchers to believe that machines will most likely
never be able to do every task for a job, and even in the most advanced scenarios, will need the
help of humans. This goes back to Butler’s point in Darwin Amongst the Machines, where AI
The third basic tendency, “Automation also complements labor,” is like it sounds,
whenever automation enters a workplace, in most cases the introduction of automation will make
the human an asset. The researchers exemplified this point through examples such as the “IT
era,” where calculations for budgets, running diagnostics on machines, or just storing records
were made easier, but produced more work to get done. A simpler example of how automation
complemented the labor, was with the assembly line, where the human workers could work
Britannica)
The fourth basic tendency, “Automation can increase demand, creating jobs,” basically
going back to the idea discussed in the first tendency. This tendency follows the supply and
demand mentality, trying to find an equilibrium, where the supply and demand are equal. This
idea could be shown through the assembly line example earlier, where a car company like Ford,
needed to make cars faster to meet the upcoming demand and lower their prices to make them
the area of job creation. This focuses on modifying jobs to do other tasks or creating new jobs
with the assistance of technology. The example the Muro’s team comes up with is that of an
ATM, where simple transactions could be made. The case they made was that with the
establishment of the ATM seeing an increase in bank tellers. This increase was caused by the
teller’s need to do simple transactions that were filled by ATMs, so they could focus on other
aspects of banking services, that needed direct contact with a client. This change was noted to
have an effect of raising the standards of education required for the job, but ultimately increased
productivity and job opportunity. This tendency is what brings together the tendencies together,
creating new jobs to fill the places where automation has taken over, whether they be entirely
The sixth and final tendency, “Technological possibility is not the same as technological
reality,” is the main point of this paper. This tendency is straight forward in telling what it’s
about. There are many factors when trying to consider the possibilities of AI or automation in
general. Some of the possible factors to consider that Muro’s team included are: “labor
competition, regulatory and social barriers, and institutional factors, among others.” This
basically all adding back up to how humans must allow AI to be a bad influence. Meaning that
humans have the control to regulate new technology and must do so to keep humanity
prosperous.
These six basic tendencies are all beliefs coming from Muro and his colleagues found.
The six come together to conclude that the “AI era” will be a repeat of the past. The “IT era,”
was found to at first eliminate a bunch of jobs due to the introduction of computers, but as the
personal computer was being developed, more jobs were created in both information
technologies and manufacturing. They easily replaced jobs with new ones, even within the same
field. This group of researchers believes that the same pattern may happen again, meaning that
there will be an initial loss of jobs through replacement, but will then eventually rebound
When relating AI to manufacturing, Tanya Anandan, a contract writer has the newest
scoop on the field. Pieter Abbeel, an electrical engineering and computer science professor at UC
Berkeley has come up with a solution that may bring AI a step closer to being sentient. Abbeel
has discovered with his team, how to teach AI using virtual reality. Humans pilot the robots,
programming them through action, then after getting a feel of what the AI is supposed to do, it
tries to replicate it. This is a slow process now, so humans still win out when compared to these
AI, in terms of how fast we learn and how fast we can move. This breakthrough though has led
AI to a place it hasn’t really been before, which is truly learning in the physical world. This
could mean that robots and AI may come sooner than expected, but I still believe that with how
slow they are now and the regulations mentioned in the Future of Life Institute’s research, it’ll be
a long while before AI can replace a person much less a part of the workplace.
Devices like “Alexa”, Amazon’s smart speaker, like “Deep Blue” are not true AI like this
one could be. Though they are classified as AI today, that is but a changing of terms as discussed
by Anandan. She believes that today’s “AI” or “smart devices,” before now, have mostly been
complex computers, storing information about the user or being able to do multiple functions.
She believes that society has begun to modify the terminology to address changing times, just
like how robots were popular, AI has begun to be the next big term. This leads to a lot of
misconceptions about what AI really is, true AI is something closer to Turing’s idea. The work
that Abbeel and his team have been working on is towards real AI.
A commonality throughout each of these sources is the fact that people must be careful in
their actions and prepare themselves for the unknown. When I mentioned that the sixth point of
Muro and his fellow researchers’ gathering and analysis of data on the cycle of AI, they also
issued this same warning. That AI is still unpredictable in its impact. This is where connecting
the Future of Life Institute’s research and the research done at Brookings is important. There is
no mistake that underestimating AI and developing it to the point where it surpasses humans as a
sentient lifeform would be dangerous. As it stands now, though it seems that the destructive AI
that we have come to know is but a mere fantasy at this point. The lack of an AI passing the
“Turing test” and the desire not to implement AI until proper safety factors are introduced, leads
I believe that with how AI is now, there is not much to worry about in terms of job loss. I
think I’ll take Gil Press’ approach to how to handle this information, taking the apocalyptic
scenario with a grain of salt for now, and just work on pursuing higher education. This way, if AI
starts filling more jobs, I’d be in a position that wouldn’t affect me as much as others, as well as
Engineering, vol. 65, no. 4, CFE Media LLC, Apr. 2018, pp. 28–31.
https://search.proquest.com/docview/2131580991/fulltext/FCED7E540F634441PQ/1?acc
https://futureoflife.org/background/benefits-risks-of-artificial-intelligence/?cn-
Butler, Samuel. “Darwin Among the Machines,” A First Year in Canterbury Settlement with
Other Early Essays. Victoria University of Wellington. London, 1914. pp. 180-185.
www.nzetc.victoria.ac.nz/tm/scholarly/tei-ButFir-t1-g1-t1-g1-t4-body.html. Accessed
Muro Mark, et al. Automation and Artificial Intelligence: How Machines are Affecting People
and Places. With contributions from Ian Hathaway, Metropolitan Policy Program at
Press, Gil. “Is AI Going to be a Jobs Killer? New Reports about the Future of Work.” Forbes.
going-to-be-a-jobs-killer-new-reports-about-the-future-of-work/#39f9d2dbafb2.
2019.