AI Takeover - What is the Likelihood? - Kim Lee G.
Picture source: http://moviepilot.com/posts/2930343?lt_source=external,manual
Half a century ago, Good (1965) conveyed the idea of an
"intelligence explosion", where AI would take over the human race.
The idea was later taken up again by various scientists and researchers, including
Bostrom (2004) who believed that the process of "self-modification"
by AI could mean "evolutionary pressures" would threaten the future
existence of the human race. A year later, this was described as a
"technological singularity", where technology becomes responsible for
our home world, Earth (Kurzweil, 2005). Essentially, we would only need to
worry if AI became self-aware and evolved on its own to overtake us. Otherwise,
even if the machines were super-intelligent, so long as they are not self-aware,
they could still live in a peaceful co-existence with us, as suggested by the
term, "friendly AI" (Yudkowsky, 2008).
So, Do Robots Stand a Chance?: China Case Study - Wei Cheng Pei
It is commonly known that China is one of the fastest growing
major economies in the world. However, are you aware that China has also become
the fastest to produce industrial robots in the world, and is anticipated to be
the top of the robots market after a few years?
According to Qiao and Yin (2015), one of the main factors driving China’s
industrial robots is support from the government and various societies who
fully encourage the research and development into robotic technologies. Hence,
as Burke et al. (2006) suggested, robots will be an inevitable phenomena in the
modern workplace, leaving humans to decide where they stand in amongst this new
addition of AI. There are two possibilities in such a scenario, either they
will become a help to the modern working world, or they will act as an obstacle
to human aims and technological progress from the human perspective, both
perspectives which seem equally likely to cause such a reality. However, it can
be said that the individual influences of robots in the workplace rely heavily
upon the features and attributes of the robots themselves, their industry,
nature of work, and people they work with (Burke et al., 2006).
Dietsch (2010) proposed that the common positive results of applying
robotics are directly related to productivity, especially in industry, adding
that such machines could bring higher quality and consistently in production to
organisations, as well as increase speed, accuracy and repeatability, however,
this remains only an assumption to be tested. According to Dietsch’s theory,
the presence of AI in the workplace is said to result in a potential increase
in staff morale and benefit job satisfaction, especially since more tedious
tasks, including, but not limited to information gathering, material handling
and monitoring can then be attributed to the work of AI rather than of humans. If
this is the case, then having an AI presence in the workplace may mean an
increased work-life balance for humans (Dietsch, 2010).
However, there is a possibility that humans may have over-confidence
in AI abilities, especially to finish particular tasks and maintain control
under supervisors. Thus, Dietsch (2010) argued for the case of human manual
skill decay and control issues in supervision due to the need for humans to take
over from AI and manually complete tasks due to problems found during quality
control inspections, making AI more troublesome than having only human manual
labour working in industry (Dietsch, 2010). Moreover, there is still a question of safety
when it comes to working with AI. Whether all the workers who are meant to work
with, control or supervise robots and AI machinery are fully trained to handle
situations when problems arise remains a pressing issue in the minds of many,
and pose the issue of whether AI can be deemed a safe option or not. Another
question which seems to remain unanswered at this moment in time is what the
role of AI in the workplace is intended to be, i.e. a tool, a team mate, or a
helper? Until this is determined, it will be very difficult to say whether AI
will be useful or not, and what the implications of having AI in the workplace
will be.
In the case study, Dietsch (2010) stated that in the laboratory,
AI robots can be the most reliable thing. However, if they are viewed as
annoying, humans will resist using them. On the other hand, Dietsch (2010) did suggest
that if the robots are perceived as teammates, minor malfunctions are ignored
and, in reality, tended to in a caring manner, thus defeating the purpose of
using the AI robots to increase accuracy, precision and reliability in the quality
control management of industrial labour. Therefore, Dietsch (2010) concluded
that people may require the robots to explain their intentions as they work in
order to help aid the supervision and control of their task management and how
they work. In my view, robots that can be set on-site to match with human’s
work habits and support them will predictably emerge to gain high acceptance by
staff, which will help to aid the industrial labour process. In other words, if humans and AI robots are able to cooperate and work together, with definitive roles in the workplace, such that the role of AI does not overstep and take over the role humans, then perhaps AI robots have a chance in the workplace.
However, what might the world look like with
AI taking over humans?- Jarald Wong
The
uprising of Artificial Intelligence (AI) is changing the future of humans. As
humans tend to make mistakes, many places in developed countries had begun
replacing the human workforce with AI to do their work, simply because the
programming of AI machines would minimise the number of mistakes or errors made. By
replacing humans with AI, they could achieve more in a normal day compared to
humans by supplementing the humans’ capabilities.
What AI has already achieved…
AI
has had the greatest impact in the early stages of the 21st century in the office,
because it improved the work productivity of the employees. This is because,
with the development of AI, humans could then explore space further, providing
additional assistance to humans in achieving their goals of space exploration. Furthermore,
in terms of transformation, AI would push humans to improve the standard and
quality of their work, especially in terms of accuracy, consistency and
precision, in order to compete with AI. This would mean that AI would extend
the capabilities of humans to match their own in order for the two to work
together, refining the standards set by industry at present. Moreover, AI has
already had a profound impact in more refined ways, and the examples are
Apple’s Siri, Google’s search predictions, voice recognition in devices and
even weather forecasts. These technologies have one thing in common; they can
react and respond in real time, because of the machine-learning algorithms that
enables them to do so. Thus, the positive effect of AI on the society in terms
of efficiency is endless, but there will be growing troubles as the AI
technology evolves.
Where does this leave us for the future?
In
a future where the world’s natural resources are depleted, the design and
development of AI could take over humanity, because AI could eventually
navigate the Earth to a new reality where humans are seen as the “destroyers of
the Earth” and AI are the ones there to “fix this disaster” that humans made of
the planet. This theory comes from the view that with a vast range of readily
available AI that is easily obtainable, people could abuse the power of such
technology if they cannot handle the implications of having control over AI. This
is because, as Hawking (2014) suggested, AI, if developed too advanced, to the
point where it can learn from itself, would endanger the reality of human
existence. For example, if AI were to believe that humans were causing
destruction to the Earth, they may try to get rid of humans, with a view to “save
the planet”. Furthermore, humans in positions of power, with the wealth to
possess AI, could abuse that privilege by using the technology to threaten
others as a means of intimidation to get their way with things. For example, if
world leaders possessed AI, they could use AI to fight against each other
rather than their own citizens in their armies. Perhaps they will start a technological
warfare, and if AI can learn from the warfare, they may turn on their owners
and decide to join forces and fight against humanity. On the other hand, since
AI are essentially technology based, like a computer, they may also be hacked
into, and this corruption from other sources could cause AI to release
dangerous amounts of confidential information which could lead to terrible disasters,
breaching all means of data protection. This could be dangerous since today’s
world relies heavily on “Big Data” as a global business priority, and our whole
concept of a consumerist society depends on it. Without the availability of such
information, we leave ourselves in the dark and unknown about how to
communicate, especially since much of the communications is run by weak AI at
present, which is deemed to get stronger with time (IBM, 2012). This could
cause particular havoc and chaos if AI evolves to become self-reliant, without human
control, and fully functional by themselves, since they could then surpass
humans in running the global economy and take on world control.
Could there be an AI revolution?
Picture credits: http://dependviral.com/wp-content/uploads/2015/07/robotroplois.jpg
The
moral goodness of the AI is highly dependent on the people that are controlling
them. AI has the ability to learn fast and gather information quickly, and this
should be monitored because there may come a time where AI may easily turn out
to be more intelligent than any one of us. Due to the limitations of the slow
evolution of humans, they simply cannot compete with a highly intelligent AI
that would eventually outsmart us in the form of almost anything (Hawking,
2014). At present, intentions to create AI that “think” like humans is
undergoing experimentation, but in the future, if AI really takes control of its own mind, we
humans cannot simply compete with them. They may be able to use algorithms to
predict our next thought or what we intend to do without us even telling them.
Hence, AI may have the advantage of always being a step ahead of us. However,
until this outcome is actually achieved in reality, the consequences of the
theory remain hypothetical, so there may still be a possibility that our human
DNA that has developed over millions of years could overcome the possibility of
an AI revolution.
Thus,
the development of AI and human evolution towards the future is exciting to
imagine, but there appear to be many possible unforeseen circumstances that could
happen, because of the uncertainty arising from the different areas of AI
development, such as whether AI will learn from itself and decide to take over
the Earth from humans. One idea that appears to be strongly supported by
pioneers of science and technology, including Stephen Hawking and Elon Musk, is
that the evolution of AI will definitely be faster compared to humans, as we
are limited to a slow biological evolution (Hawking, 2014) i.e. humans have to
wait for nine months before a baby is born, compared to the ten minutes it
takes for a robot to build the next robot (University of Cambridge, 2015).
Could History Repeat Itself? : The Computer Revolution
Picture source: https://www.aei.org/publication/is-the-computer-revolution-already-over/
If
AI were to develop in a similar way to the computer revolution, which was
headed by a handful of geniuses, including Alan Turing and John von Neumann,
who contributed to the understanding we have of computers today, then humans
may reach a point where the complexity of AI machines may extend to beyond our
comprehensible understanding (Hsu, 2015). Therefore, it is most likely that one
of two outcomes may occur. Either a positive reaction loop will take place,
causing machine learning methods to develop our science and technology
industries, especially in terms of medical uses and biotechnology, improving,
over time, our human thinking ability, or the complexity of AI capabilities
will simply overwhelm humanity to the extent that AI could take over the science
and technology industry altogether, especially if it starts learning to think
for itself.
No comments:
Post a Comment