Uploaded by edikenwagbara

CPSC Final

advertisement
Eze Dike
2259 Words
Final
Professor David Glertner
December 15th, 2021
Final Assignment
1. Why is the Question can computers have minds important?
Over the course of history humans have been able to consistently create increasingly innovative technology.
However, no leap in technological capability has been greater than those seen during the Industrial Revolution
and the three that came after. Since Charles Babbage conceived the first automatic digital computer (the
analytical engine) during the mid 1830s, humans have wondered and speculated on how far these machines
would develop. The automated capabilities of the first computers loosely resembled those of the mind. Due to
their ability to work independent of human input, once programmed: many famous thinkers expressed their
concern as to the potential of these machines. They thought machines would eventually be capable of thinking
as humans do. Therefore, the idea of conscious machines is closely tied to the idea of machines overcoming
humans. Such an idea is terrifying for some. Are we building the very machines that will be the demise of
humanity? This is the concern many thinkers share. And in order to answer such a question you must answer
several other questions. Can machines exceed human abilities? Can they be compared to humans? Can
computers have minds? If computers had the potential of achieving high-level cognitive ability synonymous
with the conscious mind, then machines would be capable of replicating, attaining and surpassing humans in all
the intrinsically human attributes. At that moment artificial intelligence would shift from something that
should be nurtured to something that must be feared.
2. Why is it a difficult question to answer?
Even though the question of conscious machines is an important one, it has historically been a difficult one to
answer. This is likely due to the controversial and paradoxical nature of the question. On the one hand it is
controversial since our understanding of the brain, and specifically consciousness is limited. So much of the
discussion on computer replication of the brain’s information processing is based in speculation rather than
sound argument based in fact. We can only really base our opinion based on what we observe from the
behaviour of computers and what we know about their inner workings. But behaviour is not necessarily
synonymous with being, and we built them but do not live inside them. So even if they act like humans does not
Eze Dike
2259 Words
Final
Professor David Glertner
December 15th, 2021
necessarily mean they are human, and even though we know how they are made we do not really know how they
experience the world. On the other hand, for a machine to be conscious would mean that it is no longer a
machine. If a machine was capable of more than it was built and programmed for it could no longer be
considered as that, so the nature of the question is also quite absurd.
3. Choose either a) or b), depending on your point of view
-
b) Computers can’t have minds, yet some thinkers (computationalists) believe they
can. Explain why they say so.
As mentioned earlier, artificial intelligence has come a very long way since its inception. Today, humans are able
to program computers to do things that were once unimaginable. This is due to the basic building blocks of
computers, which functionally resemble those of the brain very closely. In The Perceptron: A Probabilistic Model
for information storage and organization in the brain, Frank Rosenblatt writes “During the last few decades, the
development of symbolic logic, digital computers, and switching theory has impressed many theorists with the
functional similarity between a neuron and the simple on-off units of which computers are constructed ". The
possibilities that stem from these very similarities has led many theorists to believe that there will come a time
when humans will build machines, and equip them with the full set of capabilities of the human brain, through
software programing. This is wrong. Though many parallels can be drawn between the computer and the brain,
there are two major elements of cognition that are essential to what makes us human, which computers lack. No
matter how much programming is done to the software, it will not give a computer the ability to think and
emote.
Humans have come to believe that thinking is the official measure which differentiates humans and all
other beings. Renee Descartes famously wrote, “I think therefore I am,” meaning that someone who could
engage in critical thought would have to be a conscious living being to do so. With this in mind, it is important
to consider that there are important philosophical implications to the question of thinking machines, in
addition to the empirical ones. In his paper, “Can Machines Think,” Alan Turing, famously stated that he
believed there would come a time where machines would be able to think like humans. His statements are quite
optimistic seeing as humans don’t have a clue how to go about combining technological advancements with the
Eze Dike
2259 Words
Final
Professor David Glertner
December 15th, 2021
concept of free thought. In order to speculate on whether or not computers will ever be able to think for
themselves, as humans do, we must deeply reflect on human nature.
Though much of the information vital to thought is processed by the brain subconsciously, humans are
still conscious and thinking. We can think because we are conscious, not the reverse. Alan Turing asked himself
whether machines were able to do the same, and devised a way to test it named the Turing Test. In short, A
computer might be able to fool an unsuspecting human into believing that it is conscious, thinks and feels
emotion, but the truth is that the success rate of the test is only high when computers answer the human
questions with a “Yes” or “No” answer. Computers had a much more difficult time fooling the unsuspecting
human when the test required conversational answers, therefore, the test does not prove much. John Searle
argued that passing the Turing Test does not determine intelligence comparable to humans. Passing the Turing
test only demonstrates a computer’s ability to imitate human behavior based on the way it was programmed.
Programs like ELIZA have the ability to do so by manipulating symbols it does not understand fully.
Additionally, machines lacking the ability to feel emotion is a very important consideration. Emotions
are an essential aspect of human nature that thinkers, such as Turing, fail to take into account. As mentioned
above, the Turing Test is really only meant to test the imposter’s ability to deceive not anything else. The test was
adapted from a game where a person was meant to differentiate between a hidden man and woman, solely
through script-based conversation. So the test is truly only testing whether a computer can give the impression
of being a human rather than actually possessing human cognition. This is exactly where Paul Ziff finds
contention with much of Turing’s work. It fails to factor emotion into the broader thinking of human beings,
therefore it contains no real metrics for emotional intelligence. In Mentality in Machines, D. M. Makay points
out that any test based on the observable activity of a human can be passed by automata. So the Turing test - and
any other test which measures based on observable behaviors - is not at all an appropriate test to calculate a
machine’s intrinsic likeness to a human being. Ziff goes even further by arguing that robots are simply
mechanisms that can not - and will not - ever “feel” as living organisms do. And machines are sometimes capable
of fooling humans because they are specifically designed by humans to fool other humans. “ No robot would
act tired,” he writes. Then follows with, "Because a robot could only act like a robot programmed to act like a
Eze Dike
2259 Words
Final
Professor David Glertner
December 15th, 2021
tired man.” Really machines can only be as human as they were programmed to be, and such a reality is
inhuman, in and of itself. For machines to truly eventually be synonymous with cognitive functioning, humans
would have to figure out how to program consciousness into software. Consciousness is often defined as the
ability to know and perceive. It is through consciousness that one becomes aware of the implications of their
actions, as well as themselves in the moment. To understand the meaning behind actions, memories are
necessary. Memories give humans the ability to attach context from the past to things they will do in the present.
This ability is an important part of language and morality, which ultimately help make us human.
Like humans, machines can store memories. But there are some important distinctions that must be
made when comparing the brain to a computer. Though machines can recall memories at a much better rate
than the human brain in much greater detail; the Neural Network that stores the memories within its Neurons
must do so through an algorithm that is externally programmed. The individual neurons of the network do not
know for what purpose they are being programmed. That is where computers differ from the human brain, and
it is highly unlikely this difference will ever change. The human brain contains different independent
components, some that operate on the purely abstract, some that operate on the specifically defined, and some
that operate on both. When the brain is given multiple tasks, its neurons themselves adjust for how to solve or
how to keep more than one mapping amongst the so-called “infinite neurons in the brain.” It is as though
evolution has allowed human neurons to come pre-programmed to process information dynamically. This
“pre-programming of the human brain” is what allows humans the ability to exercise free-will, which computers
can’t.
According to the Philosopher Adam Smith, “free will” is “being able to accept responsibility.”(Smith)
Even if computer scientists were able to figure out how to equip a computer with software that enables random
and arbitrary thoughts, whether it is through a set of random algorithms or some other means, and this
computer processed information through simulated neurological pathways similar to a human brain, which
enables it to think in a non-cyclical pattern. It would just be an imitation of the human brain and not its own
original mind. It would still lack free will, and base its actions (output) on either syntax or an input. This is why
John Searle brings up language in “ The Chinese Room” experiment. Computers will not ever really be able to
Eze Dike
2259 Words
Final
Professor David Glertner
December 15th, 2021
truly understand language, which is why Searle uses the experiment to oppose Turing’s views. A language
consists of any number of valid sentences. Two things are necessary to assure the validity of any sentence: proper
syntax and semantics. While syntax simply involves the grammatical structure of said sentence; semantics alludes
to the meaning of the vocabulary symbols placed within the grammatical structure of the sentence. A computer
will never truly understand why the sentence “apples sing gracefully” has no meaning, even if it is programmed
to refuse the sentence. Semantics are out of the reach of computers, and will be forever.
In summary, if machines had the potential to replicate, attain and surpass the human level of cognitive
ability synonymous with the human conscious mind they would be feared. That is the cause of the interest so
many have in their mental ability. We can only really attempt to understand the cognitive ability of machines
through their inner workings and actions, but neither helps us to truly know if they are capable of human-like
cognition. So the discussion is left to opinion and speculation. And it is my opinion that no matter how much
innovation the field of programming sees, it will not give a computer the ability to think and emote. Humans
will never be able to create a mind quite like our own and there is a large amount of evidence in support of this.
Computers have emerged that are even able to learn on a curve similarly to the human brain. Even so, for an
engineered inorganic machine to surpass the innate capabilities of the human mind means that such a
mechanism could no longer be considered as a mere machine. As mentioned earlier, the “pre-programming of
the human brain” is what allows humans the ability to exercise free-will, which computers can’t. Therefore, for a
machine to engage in critical thought it would have to be a conscious living being. But that is not the case,
machines are simply inanimate objects, designed to carry out the tasks we assign to them. That is why the Turing
Test is a poor test and robots will never truly be able to comprehend semantics.
Eze Dike
2259 Words
Final
Professor David Glertner
December 15th, 2021
Citations
Rosenblatt, Frank. "The perceptron: a probabilistic model for information storage and
organization in the brain." Psychological review 65, no. 6 (1958): 386.
Smith, Adam. “Theory of Moral Sentiments”
Turing, Alan. “Can Machines Think”
Wisdom, J. O., et al. “Symposium: Mentality in Machines.” Proceedings of the
Aristotelian Society, Supplementary Volumes, vol. 26, [Aristotelian Society, Wiley], 1952,
pp. 1–86, http://www.jstor.org/stable/4106628.
Ziff, Paul. “The Feelings of Robots.” Analysis, vol. 19, no. 3, [Analysis Committee,
Oxford University Press], 1959, pp. 64–68, https://doi.org/10.2307/3326566.
Download