6

IBM's Watson computer is known for winning the Jeopardy quiz show against the top human players. It has also been put to cancer research.

Watson works by analyzing large chunks of texts to come up with answers, not relying on structured data such as tables etc. When winning Jeopardy it relied on Wikipedia, among other things, and while doing cancer research, it relied on peer reviewed scientific articles.

Would it be possible for Watson to answer philosophical questions if it were given a lot of philosophical texts as its training material?

  • 1
    to put it briefly - its like confusing the index of a book with who wrote it. – Mozibur Ullah May 1 '14 at 13:28
  • 4
    But Watson can edit & improve your question, apparently ... – rolfl May 1 '14 at 13:38
  • 1
    What fairly simple neural network can do youtu.be/vShMxxqtDDs?t=53m19s ;) – Lucas May 1 '14 at 15:53
7

No. Watson is just using an algorithm to look up answers to questions that are already definitively known. (What is X correlated with: Y or Z?)

Philosophical questions aren't like that. "What is knowledge?" Watson would spit you back something like, "Knowledge is justified true belief, where the connection between justification and belief isn't just lucky" or something like that. Ok fine. But Watson is just going to be reporting the conclusions of arguments philosophers have already invented.

We're nowhere near having an AI that is able to create novel philosophical arguments for substantive theses.

  • 1
    Missed something in last sentence between having and that. – Asphir Dom May 1 '14 at 12:15
  • 1
    Arguably, many philosophers also are incapable of coming up with novel arguments :-) – James Kingsbery May 1 '14 at 15:55
  • @AsphirDom, Fixed. – shane May 1 '14 at 17:00
  • 1
    @JamesKingsbery, agreed. But the fact that most humans suck at philosophy, doesn't mean that Watson would do better. – shane May 1 '14 at 17:01
  • 2
    @shane Excuse me? – user3164 May 1 '14 at 17:01
3

Yes, some of them. Watson, or something like it, could certainly answer the majority of the questions asked here at Philosophy.SE

This is the highest scoring answer on this site, and it "works" by recognising the asker's topic and then providing two Jepody-like answers: "what is Intuitionism?", "what is Platonism?" with their definitions. It's exactly the kind of problem an expert system is good at solving.

An expert system will probably perform quite well with the scholarship of philosophy, knowing who said what and putting ideas in the context of others. This is a big part of philosophy, more so than most other academic disciplines. But a Watson-like AI would probably struggle to articulate philosophy that humans would consider truly novel.

Then again, as Mark Twain is said to have said:

There is no such thing as a new idea. It is impossible. We simply take a lot of old ideas and put them into a sort of mental kaleidoscope. We give them a turn and they make new and curious combinations. We keep on turning and making new combinations indefinitely; but they are the same old pieces of colored glass that have been in use through all the ages.

A Watson-like AI might surprise us!

  • 1
    +1, but, looking at your comment to the question, which was quite funny, I suppose early versions might excel in postmodern philosophy. – user3164 May 1 '14 at 17:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.