AI researchers think 'Rascals' can pass Turing test "We are building a knowledge base that corresponds to all of the relevant background for our synthetic character--where he went to school, what his family is like, and so on,"
"We are building a knowledge base that corresponds to all of the relevant background for our synthetic character--where he went to school, what his family is like, and so on,"
- Having a big knowledge base is not a sufficient condition to pass the Turing Test. To discover this, google search "how fast can a horse run" followed by "how fast can a dead horse run".
I'll be interested to see the transcript from this attempt to pass the TT. I wonder how long it will be before the algorithm makes a "dead horse" mistake...
Personally, I don't think it'll pass the Turing Test. I think no AI will ever pass the Turning Test before the Turing Test itself is dismissed for being irrelevant. What I mean is, when AI is advanced to the level that it surpasses human intelligence, we'll be "surprised" to find out that AI didn't turn out to think like us; thus, Turing Test becomes useless because it'll show how accurate the AI can imitate a human rather than how truly smart it is.
I did search for how fast can a dead horse run but I missed the point you are trying to make. Though I agree with you that knowledge base is not everything.
the point is that if you search for "how fast can a horse run". the first hit on google is a link to the correct answer - 30 miles per hour or whatever.
but if you search for "how fast can a dead horse run", nothing on the first page of hits tells you the correct answer - 0 miles per hour. this is because google doesn't have common sense.
I guess they're further ahead than I thought. They're actually planning to do the test this fall. I would definitely be interested in attending if it's open to the public.
- Having a big knowledge base is not a sufficient condition to pass the Turing Test. To discover this, google search "how fast can a horse run" followed by "how fast can a dead horse run".
I'll be interested to see the transcript from this attempt to pass the TT. I wonder how long it will be before the algorithm makes a "dead horse" mistake...
I did search for how fast can a dead horse run but I missed the point you are trying to make. Though I agree with you that knowledge base is not everything.
but if you search for "how fast can a dead horse run", nothing on the first page of hits tells you the correct answer - 0 miles per hour. this is because google doesn't have common sense.