If a computer can pass the Turing test, it means that if a person talks to a computer from a remote node, that person won't be able to figure out whether or not who he's talking to is another person or AI.
When will computers pass the Turing test?
- Thread starter apendrapew
- Start date
- Status
I'd give it about 10-15 years when computers start behaving more like human brains, ie, when they become analog and crunch calculations in a massively parallel way.
eburacum45
Valued Senior Member
The problem with the Turing test is that some people are easily fooled; there are already many instsnces of computers fooling people into thinking they are human; this is usually because the humans are inexperienced in this field...
what a real turing test needs to do is fool an expert in artificial intelligence; what this will prove is open to debate.
Tomorrow - Twenty years- a hundred. Who knows.
__________________
SF worldbuilding at
http://www.orionsarm.com/main.html
what a real turing test needs to do is fool an expert in artificial intelligence; what this will prove is open to debate.
Tomorrow - Twenty years- a hundred. Who knows.
__________________
SF worldbuilding at
http://www.orionsarm.com/main.html
Uhh, also some people are easier to emulate. I have met people here on the internet that would be pretty easy to emulate. I think this is the problem with the Turing test: It is very fuzzy; what kind of person is the computer supposed to emulate and what skills is the evaluating person supposed to have?
Hans
Hans
Okay, say the program that you're talking to is pre-programmed like all of us to gain information from the world and hence form its own "personality". So it would have a perfect memory and it would have the choice to be anything it wanted to be. Given that it has enough stimulii to learn from, which is an insane amount if it's free to surf the internet, it could be pretty damn knowledgeable. Say the evaluating person were you. I know right now, I can find out whether or not I'm talking with an AI. But uh, what's your guess?
I think it would be easier to emulate emotions. Not that the computer will HAVE emotions, but it will seem to. I once made a program with a nasty temper (as part of a game), that did not seem too difficult.
I can't even imagine how I would emulate humor, OK, I could make it crack jokes from some inexhaustible supply, and puns might be practicable too, but that kind if humor that makes us make a screwd remark, and return one .... no.
Hans
I can't even imagine how I would emulate humor, OK, I could make it crack jokes from some inexhaustible supply, and puns might be practicable too, but that kind if humor that makes us make a screwd remark, and return one .... no.
Hans
fugazzi007
Registered Member
I think the point is not to emulate human behavior but to create original ideas given a set of circumstances. The ability to have this sort of semi cognizance, such that the program is able to create a non-canned response, will be a turning point in the way we view computers.
On the other hand, it's easy to talk to real people online and get 100% canned responses. Maybe an original-thinking program would be an obvious pointer as to its lack of humanity.
On the other hand, it's easy to talk to real people online and get 100% canned responses. Maybe an original-thinking program would be an obvious pointer as to its lack of humanity.
- Status