Hritik complex issues without any human input. The ability

 Hritik Panchasara

Professor Stout

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now


1 November 2017

Are the misconceptions surrounding Artificial Intelligence hampering its
own progress?

Are the misconceptions surrounding Artificial Intelligence hampering its
own progress?


Are careers like financial analysts or telemarketing necessary for us
humans to labor at? Could the Greek mythological figures be simulated and
brought to life using a form of superintelligence? Our technology is on a path
of such high magnitude that it could shape our future for the better. The
program JARVIS from the movie Iron Man is a highly advanced computer artificial
intelligence that managed everything that was related to technology for the
protagonist. The fact that something inorganic can be of such high value goes
to predict the future of our own technological race. Artificial intelligence is
defined as a subfield of computer science wherein computers can perform tasks
for humans that we would normally think of as intelligent or challenging. Envision
a future where computers and machines can carry out our daily human tasks at
ease and solve complex issues without any human input. The ability to invent
intelligent machines has fascinated humans since the ancient times. Researchers
are creating systems and programs that could mimic human thoughts and try doing
things that humans could do, but is it here that they got it wrong? Humans have
always been good at defining problems but not solving them. Machines, on the
other hand, are polar opposites, where their computational power helps them
solve almost any problem, but not define them. It goes to show how these two
aspects are interdependent on each other and why we are looking forward to the
invention of superintelligence. But issues like creationism and negative
typecasting beg the question of whether the misconceptions surrounding
superintelligence are hampering its own progress. A few scholars like Pei Wang
focus on the dynamics of a working model and the inaccuracies in it. But
scholars like Yoav Yigael question the emulation of human-like characteristics
and abilities into machines. This research paper will focus on the various,
incorrect approaches towards harnessing this technology,  the consequences
that are being derived from it, and the solutions that could probably be
focused on.

One of the main issues surrounding artificial intelligence is the fact
that global leaders have an illusion of what it is supposed to be. They
constantly try and emulate human beings in machines when that was not the goal
of the technology since its inception. Take the wheel as an example. The wheel
was supposed to augment human capacity for transportation and it successfully
paved the way for countless other inventions. In the same way, artificial
intelligence was meant to augment to our cognizance and help us function in a
better manner; to solve problems that we could only define. The most common
trend is the creation of humanoids like Hanson Robotics’ Sophia. It is
an amalgamation of artificial intelligence and cutting-edge analytical software
for peak performance as a “question answering” machine and it’s more so an “Andro-humanoid” robot. Elsewhere, IBM
and its drive to replicate human nature was not only unsuccessful, but also has
been causing a financial burden on the company. IBM simply tried too hard to
push Watson into everything ranging from recipes to health care and it resulted
in a declining revenue for 5 years now. Hence, it alludes to the
misappropriation of resources by feeding research into pointless products and
avenues for a largely versatile product.

Artificial intelligence used to be a problem solving machine where
commands were entered in an parameters box. Human programmers would
painstakingly handcraft knowledge items that would then be compiled into expert
systems. These served to be brittle to a certain extent and could not be
scaled. Since then, a quantum leap has changed the field of artificial
intelligence. This pioneered the idea of superintelligence but somewhere along
the way it has been grossly misunderstood. Machine learning, is what has
revolutionised how we make and train AI. Earlier knowledge items and structures
were pre-defined by manual programming, now machine learning enables us to produce
algorithms that learn, from unprocessed perceptual data. This process can be
likened to how human infants learn. Is it possible for us to take a system of
devices interconnected and co-dependent, and process their data in meaningful
ways, to pre-empt their shortcomings and avoid errors? Yes. Is it possible for
us to build a machine so adept in our languages that we can converse with it
like we do with each other? Yes. Can we build into our computer’s a sense of
vision that enables them to recognize objects? Yes. Can we build a system that
can learn from its errors? Yes. Can we build systems that have a theory of mind?
This can be done using neural nets. Can we build systems that have a moral and proper
foundation? This we are still learning. A.I. is still miles from having the
same potent, pan-domain ability to study and plot as a human being has. Humans
have a neurological advantage in this case, the power of which we yet do not
know how to replicate in machines.

Ever since AI’s inception this question has been asked, is it something
to fear? Every advancement in technology draws upon itself some apprehension. The
invention of the Television was criticised as people complained that it would
make the working class procrastinate and make them dull. On the creation of
E-mail, society grieved that the personal touch and formality of a letter would
be lost. When the Internet become pervasive, it was argued that we would lose
our ability to memorize. There is truth in all these claims, but it’s also
these very technologies that define our way of modern life, where we have taken
information exchange for granted no matter what the medium, this in turn
expanded the human experience in substantial ways. The film, “2001: A Space
Odyssey” by Stanley Kubrick personifies all the stimuli that we have come
to associate with AI, as one of the central characters is HAL 9000, an AI. HAL,
a sentient computer programmed to assist the Discovery spacecraft from the
Earth to Jupiter. It was a flawed character, as in it chose to value its
mission objective more than human life. Even though Hal’s character is rooted
in fiction, it voices mankind’s fear of being subdued by a being of superior intelligence
who is apathetic to our humanity. The AI that researchers and scientists are
trying to make today, is something that is very much along the lines of HAL,
but without its single mindedness of achieving its objective without nuance. This
is a hard engineering problem, to quote Alan Turing, “We can only see a short
distance ahead, but we can see plenty there that needs to be done.” 


To build a safe AI, we need emulate in machines how humans think, this
is a task that seems beyond impossible, but it can be broken down to three
simple axioms, the first axiom is altruism, if the AI’s only goal is to
maximize the comprehension of our objectives, and of our values. Values here
don’t mean values that are distinctly intrinsic, extrinsic or purely moral and
emotional, but a complex mixture of all of the above, as we humans aren’t
binary when it comes to our moral compasses. This actually violates Asimov’s
law that states robots have a sense of self-preservation. Whilst preserving its
existence is no longer its priority whatsoever. The second axiom is of humility.
This states the AI does not know what our human values are, so it maximizes
them, even still it does not know what they are. This ambiguity of our values
is of our advantage over here, this helps us avoid the problem of single-minded
quest of an objective, like HAL. In order to be of use, The AI has to have
rough impression of what we want. It acquires this information predominantly by
observation of our choices. The question then is what happens if the AI is ambiguous
about the objective?  It reasons
differently. It considers the scenario where we could turn it off, but only if it’s
doing something wrong. AI’s do not know what wrong is, but it reasons it does
not want to do it. In this scenario we can see the first two axioms in action.
Hence it should let the human turn it off. Statistically you can estimate the motivation
that the AI has to permit us to turn it off, and it’s directly proportional degree
of uncertainty of the objective set for the AI. When the AI is turned off, that
third axiom comes into play. It infers something about the objectives it should
be pursuing, because it infers that what it did was not right. We are factually
better off with an AI that’s designed in this way than with an AI built any
other way. The scenario above is an example which depicts what humans endeavour
to accomplish with human-compatible A.I. This third axiom draws up
apprehensions from the scientific community, because humans behave badly. A lot
of human behaviour is not only displeasing but also wrong, which means that any
AI that is based on human values will corrupt itself in the same the humans
have. What one must remember is just because the maker behave poorly doesn’t
mean the creation (AI) is going to mimic that behaviour. The fundamental goal
of these axioms was to provide nuance for why humans do what they do, and make
the choices that they make. The final goal is to allow AI to predict for any
person the outcome of all their action/choices in as accurate a manner as possible.

The bigger problem now is, how do we feed in all our values and morals
and the nuances that are associated with them into an Ai which is essentially
an inference engine at this point,  doing
this the old school way by manually defining every knowledge item would be
impractical , Instead we could leverage the power of A.I. here, we know that it
is already capable of processing raw perceptual data at blinding speeds, so we
essentially use its intelligence to help us in helping it learn what we value,
and its incentive system can be fashioned in such a way that it is incentivised
to pursue our ethics or to perform actions that it calculates we would approve
of using the three axioms stated above. In this way we tackle the difficult
problem of                Value-loading
to AI with the resources of an AI. It is possible to build such an artificial
intelligence, and since it will embody some of our morals, the fear that people
have for AI of this capacity is baffling. In the real world constructing a cognitive
system is fundamentally different than programming an outdated software-intensive
system. We do not need to program them. Programmers start to teach them. In
order to teach an AI how to play a game of chess we have it play the same game of
chess a thousand times, but in the process we also teach it how to discern a
good game from a bad game. If we want to create an AI medical assistant, we
will teach it endocrinology whilst simultaneously also fusing with it all the
complications in a person that could lead to the underlying symptoms. In technical
terms, this is called ground truth. In programming these AI, we are
therefore teaching them a sense of our morals. IN such cases, humanity must trust
an AI equally if not more as a human who is just as well-trained.

“Superintelligence” the book by the academic Nick Bostrom, he
reasons that AI could not only be dangerous to humanity but it’s very existence
one day might spell an existential crisis for all of humanity. Dr. Bostrom’s
primary dispute with AI is, that such cognitive systems  learn on digital time scales and that means
soon after their inception they will have inferred and deciphered all of human
literature, this alludes to their ravenous huger for information and there
might come a day when  eventually when it
ascertains that the objectives set for it by humanity no longer align with its
own objectives and goals. Dr. Bostrom is held in high regard by people of
immense stature such as Elon Musk and Stephen Hawking. With all due respect to
these academics and philosophers I feel that their assumptions about AI are
erroneous to an extent.  Consider the
example of HAL as stated above was only a hazard to the Discovery crew so as
far as it was in command of all features of the Discovery spacecraft. This is
where Dr. Bostrom faltered in assuming that the AI would have to have control
over all of our world. The most popular stereotype of Skynet from “The
Terminator” is a prime example of such a scenario. It was here in the movie
that a superintelligence eventually took command of mankind by turning all
machines against humanity. However, we must remember that our goal with AI was
never to build AIs that could control and harness the weather, that would
direct and manipulate the tides, which could command us whimsical and disordered
humans. Furthermore, if such an artificial intelligence existed, it would have
to compete with human economies, and thereby compete for resources with us.
Furthermore if the three principles stated above are used as guidelines in the
formulation of this omnipotent AI, then not only do we not fear this AI but we
cherish it, for it is built in our image, with our values and morals. We cannot
protect ourselves from all random acts of violence, Humans are unpredictable
and the truth is some of us are extremists, but I do not think that an AI could
ever be a weapon that an non-governmental tertiary party could ever get its
hands on, and to manufacture an AI by these parties is even more farfetched as
the mobilizing of resources and brainpower alone would raise multiple red flags
for the authorities of the world to stop whatever devious ploy to overthrow
world order in its tracks. 

Artificial Intelligence is heading into multiple directions and there is
a lack of a centralised effort for the development and advancement of this
science towards a neutral goal. Moreover, humans anthropomorphize machines and
this leads them into believing that the flaws of the maker will be heightened
in the flaws of its creation. There are some obscure problems neural cortex of
any AI and how is it that we make it conscious, what is conscience? Questions
like these need to be answered before we march onwards on our quest of an
omnipotent AI. Furthermore intricacies of the decision theory for an AI are
still primitive in its infancy so we some way to go before we figure that out.
These problems seem far to advanced and complex to tackle now, but the truth is
that the research is already underway and sooner rather than later we’ll
witness the ushering in of the era of machine intelligence.

Works Cited:

Wang, Pei. “Three Fundamental Misconceptions of Artificial
Intelligence.” Taylor and Francis Online, 13 August 2007,
Accessed 13 November 2017.

Yigael, Yoav. “Fundamental Issues in Artificial Intelligence.” Taylor
and Francis Online, 7 November 2011,
Accessed 13 November 2017.

Yudkowsky, Eliezer. “Artificial Intelligence as a Positive and Negative
Factor in Global Risk.” New York: Oxford University Press, 2008, Accessed 14 November 2017.

Hammond, Kristian. Practical Artificial Intelligence for Dummies.
John Wiley & Sons, Inc, 2015. Accessed 14 November 2017.

Bostrom, Nick. Superintelligence: Paths, Dangers, Strategies. Oxford
University Press, September 3rd 2014. Accessed 9 December 2017.