From the readings, what is artificial intelligence and how is it similar or different from what you consider to be human intelligence?
Are AlphaGo, Deep Blue, and Watson proof of the viability of artificial intelligence or are they just interesting tricks or gimmicks?
Is the Turing Test a valid measure of intelligence or is the Chinese Room a good counter argument?
Finally, could a computing system ever be considered a mind? Are humans just biological computers? What are the ethical implications are either idea?
The study of intelligence is one that has confounded nearly every approach from science to sociology. In essence, discussions on intelligence concern themselves with evaluating how a specific actor makes decisions in response to stimuli. Certain response are considered more intelligent than others, and how often those intelligent decisions are chosen determines an actor’s overall intelligence. Already subjectivity is introduced into the definition, for declaring which decisions are intelligent grows exponentially more difficult with the complexity of a situation. Evaluating the decision-making process becomes even more difficult, however, when looking more deeply into how decisions are made; the field of philosophy in particular offers logical arguments against the feasibility of evaluating intelligence. One relevant field of thought is called skepticism.
One of the most famous philosophical statements (in its full form) is Descartes’ “I doubt, therefore I think, therefore I am.” His claim is that we cannot be deluded as to our own existence: if we did not exist, then there would be no target of delusion, therefore we must exist, at the very least so we can doubt our own existence. Descartes’ conclusion reveals a central concept in the study of intelligence: the most informed perspective on a specific intelligence is the intelligence itself. The skeptic takes this conclusion and goes further, claiming that the only intelligence that we are sure exists is our own. A popular expression of this belief is the question, “What if everyone else is a robot?” There is no way to be completely sure that the people surrounding us are not deterministic facsimiles of humans, since we cannot see into their heads.
We can, however, see into the minds of artificial intelligences. But does this actually make it easier to classify and define intelligence? Many intelligence evaluation experiments choose to ignore this access: the Turing test is one of the most famous, and it uses only external signals to infer intelligence. The skeptic denies that passing the Turing test is a result with any real meaning, since there is no way to know how the behavior exhibited during the test came to be. But even if we knew, what can we conclude? Intuitively, we don’t consider a finite state machine, which simply acts on a set of rules given an input, to be an intelligence, no matter how sophisticated. We also intuitively consider ourselves to be intelligent beings, but we aren’t even sure of how we as humans come to decisions! And even if we were, it seems shortsighted to claim that only our method of decision-making qualifies as sufficiently intelligent.
Observation-based approaches such as the Turing test, while useful in order to colloquially call something “intelligent,” fall short of describing a complete definition of intelligence, as shown by the skeptic’s argument outlined above. Some argue that such a definition is not needed, but without one, we are unprepared to answer moral questions on the rights of artificial intelligences and the nature of our existence in comparison. We must investigate our own decision-making process further as well as those of any other intelligences we discover in order to build a stronger catalog of those decision-making processes we intuitively consider intelligent. Only then will we be in a position to decree what qualifies as intelligent.
Zach's Ethics Blog 