Q&A: Artificial Intelligence and Plato
Artificial Intelligence and Plato
Question
Good evening!
Today artificial intelligence has the ability to learn to make new generalizations beyond what the programmer taught it. My question is: how is this possible? After all, even the human ability to make such a generalization and analogy is based, according to Plato, on an ideal observation that seemingly only a human can perceive. I saw that the Rabbi wrote in the column that the programmer taught it this, but that would not help, since the intelligence always learns beyond what it was programmed to do.
[It seems to me that this question is difficult even according to Aristotle, since generalization is an abstraction performed by the mind. This is because, in order to compare items, there must be some criterion, but there are infinitely many items; so how can one compare and evaluate them? For example, every triangle is different, and there are countless things similar to a triangle. That is, any similarity whatsoever is based on some external criterion. In truth, seemingly this question is difficult for Aristotle himself as well, no?]
Thank you very much!
Answer
I don’t understand the question. What do you see in artificial intelligence that goes beyond what the programmer taught it?
See column 653, which was posted the day before yesterday.
To be honest, I also thought about these things and asked myself whether maybe there really is something in the world like the "active intellect" that is responsible for the intelligence of animals, human beings, and rare natural phenomena (in Epochtime magazine, in an article by Rakefet Tavor from July 2023, there is an article about whether a fourth state of matter [plasma] in the universe is actually a living creature). And if so, maybe we gave artificial intelligence the tools to create a connection with the active intellect.
It may be that I’ve gone completely overboard and this is really an idea for a Hollywood movie, but maybe Aristotle was right.