Does Spectrum people pass Touring test?

Post Reply New Topic

This is the Touring test: http://en.wikipedia.org/wiki/Turing_test

Basically is a test about Artificial Intelligence. I'm pointing out that because I see in it the same kind of "racism" as people had (have) about autism. Basically the test is as follow:

A - robot
B - Human
C - Human esaminator.

C don't know if A or B is the robot. If C can find wich one is the robot then the test is not passed.

People said to me so much time that my mind is "robotic" or that I "don't think like others" or things like that, so I asked myself: "Can I pass touring test?".

If you look at the link there is a discussion about it (not about Spectrum). Look at "Weaknesses of the test" subtitle.

P.S.
A cool quote: "planes are tested by how well they fly, not by comparing them to birds"


EDIT:
Probably I was saying it in the wrong words.

The question is: suppose that you put under a screen an NT and a AS. Then you say to the "operator": "one of them is a computer. Wich one?" Do you think we will have a 50% chance, more or less to be "labeled" as computer?

So you are asking if people are Turing complete? Being that you are not a machine, it's an invalid question. That's sort of like asking, "If an apple tree were an animal, would it walk around?"

Probably I was saying it in the wrong words.

The question is: suppose that you put under a screen an NT and a AS. Then you say to the "operator": "one of them is a computer. Wich one?" Do you think we will have a 50% chance, more or less to be "labeled" as computer?

I remember awhile ago talking to an artificial intelligence computer online that they were trying to get to pass the Turing test. The point was that the computer wants YOU to convince IT that you're human. I think I gave up; it was REALLY difficult!

Not certain about the general Turing test, but I'm certain we could likely pass the SME Turing test...

We will pass SME test probably better than NT
http://en.allexperts.com/e/s/su/subject ... g_test.htm


ALICE: http://alicebot.blogspot.com/
MR.Mind (convince him you are not a computer): http://www.mrmind.com/mrmind3

I think this t-shirt would be appropriate for many people here

Link

The problem is you can't base intelligence on things like the Turing test. It's assuming a major fallacy that your ability to communicate is related to your ability to discern concepts and ideas. An intelligent extra-terrestrial may very well have communication simpler than that of a humble toad, yet have no difficulty outpacing us in technological innovation.

We've been seeing the impact of this concept in the Raven's Progressive Matrices, and what it shows about autism when compared to standard IQ tests.

For us to determine the intelligence of anything, we need to completely ignore language, and instead sit back and watch. Does this creature use tools? Is it capable of putting things together in innovative ways? Is it capable of learning and developing abstract concepts?

If we can look at it from the perspective of intellect, rather than complex communications, we'll get a much more accurate picture. Dolphins have a language more complex than ours, yet I'd hardly consider them 'intelligent life'.

Yeah, we're not really talking about intelligence here, but "humanity"--whatever quality NTs see as "human", that chatbots and the like don't have.

I think possibly some autistics might fail the test. I might've failed it back when I was a kid, before I learned how to do conversations. The give-and-take you get when you're talking to somebody can be pretty difficult for autistics. That doesn't mean they're not human; just that they don't communicate it to others. Think of it as a sort of "false negative", like the "false positive" of a computer that passes the test.

Most chatbots now, when they make up what they'll say next, are responding just to what you said to them, by keyword most likely. They'll search their entire database for something appropriate to say next. That's not like a human (well, most humans); humans will have a conversational topic in mind, and search the things that are in that topic and most closely related, in their own framework of ideas, to whatever the person before them has said to them in the last half-hour, not just the last statement. Then there's other complications, like words with more than one meaning, sarcasm, metaphor, simile, slang...

The one thing humans are really, really good at that computers suck at is all sorts of pattern-matching. So even as a kid, before I figured out conversations, when I was still just using pattern-matching (i.e., saying what I thought matched the particular situation I was in) I was probably doing better than the best modern AIs at imitating conversation.

it's not the "Touring" test.