Slashdot Mirror
New Contestants On the Turing Test
vitamine73 writes "At 9 a.m. next Sunday, six computer programs — 'artificial conversational entities' — will answer questions posed by human volunteers at the University of Reading in a bid to become the first recognized 'thinking' machine. If any program succeeds, it is likely to be hailed as the most significant breakthrough in artificial intelligence since the IBM supercomputer Deep Blue beat world chess champion Garry Kasparov in 1997. It could also raise profound questions about whether a computer has the potential to be 'conscious' — and if humans should have the 'right' to switch it off."
and see if it complains, first. If it does, then call me back.
Are they really *thinking* or have the programmers just done some tricks to make it seem that way.
"Teaching to the test", so to speak.
If we don't have the right to switch off a conscious machine (one that passes the Turing test) does that imply we have the right to switch off a human who fails a Turing test?
The purpose of (strong) artificial intelligence isn't to trick humans somehow, it is to figure out how our mind works. What is the algorithm that powers the human brain? No one knows.
Who cares if contestants can be tricked by a computer? Who cares if some computer can calculate chess moves faster than any human? None of this helps us get closer to the real purpose of AI, which is why they call it weak AI.
Qxe4
It could also raise profound questions about whether a computer has the potential to be "conscious" -- and if humans should have the 'right' to switch it off."
Maybe in the esteemed opinion of vitamine73 it will, but if you knew anything about how artificial conversation engines were constructed, you would understand that it's anything but sentient. Right now, conversation logic is simply trick laid upon trick to stagger through passing as a human, and doesn't, at its core, contain anything remotely similar to self-aware thought.
I'm sorry. The number you have reached is imaginary. Please rotate your phone 90 degrees and try again.
...AI is an exciting subject, but the Turing test is pretty crude.
The Turing test doesn't tell you whether a machine is conscious or self-aware... All it tells you is whether or not a programmer or group of programmers created a sufficiently advanced chat-bot. So what if a machine can have "conversations" with someone? That doesn't mean that same machine could create a symphony or look at a sunset and know what makes the view beautiful.
Star Trek androids with emotion chips should stay in the realm of Star Trek, because they're surely not happening here.
It is likely to be hailed as the most significant breakthrough in artificial intelligence since the IBM supercomputer Deep Blue beat world chess champion Garry Kasparov in 1997.
I don't understand how this is a breakthrough for artificial intelligence. Deep Blue didn't "think", at least not in the way most people think when they consider artificial intelligence. It did what computers are really good at - it computed.
Deep Blue applied an evaluation mechanism specifically tuned to chess - taking the location of pieces on the board and computing a number telling it how "bad" or "good" this position was and how "bad" or "good" responses to this position would be. Granted, it took this to a depth farther than any other chess computer in history, but it was doing essentially what a small, handheld chess computer does.
Of course a computer is going to be good at computing. That doesn't mean it's thinking.
Early chess computers used AI techniques to try and cut out candidate moves. This was expensive in CPU cycles, but the thought was to get them to play chess like humans. Computer chess since AI Winter has been all about number crunching - let Moore's Law take hold and just brute force our way through the problem - evaluate deeper because we have a faster processor. This is what Deep Blue did.
If Deep Blue were true AI, then it wouldn't be limited just to chess. It's an interesting experiment in computer chess, and an interesting experiment in tuning an algorithm working against a human, and in interesting experiment in making a computer chess opening book, but a huge leap forward in AI it isn't.
If you read TFA they have a sample chat which just shows you how stupid these chat bots still are. It is extremely easy to get them to just parrot responses and then try to change the subject in completely random directions.
I have yet to see any chat bot that can figure out the line of questioning, then pick up and introduce interesting things to the conversation that are corollary to that subject. I think the only way you will get bots that will "pass" this test is to have massive databases of words, relationships between words and subjects with corresponding topics of discussion. Still, the computer won't be intelligent, it will just be reciting from its huge database of responses.
I think the type of question i'd ask these bots is something that would require them to extemporize and they'd all fail. For example: "You have two rubber ducks, what are the possible ways you could use them if you don't have a bathtub?"
Any human could reply to that with things like "i'd put them in a stream, run over them with my car, put them on a lake, in the swimming pool" etc but a computer program isn't likely to respond to that in any way that makes sense. The response i'd expect from the computer would be "You like ducks then?".
"To strive, to seek, to find, and not to yield." - Tennyson
Hey, that's not funny! My sister died that way.
Conscience is the inner voice which warns us that someone may be looking.
Now we'll see the vim vs emacs flame war. GREAT! THANKS!
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
Can you tell human from machine?
One of these conversations is between Professor Kevin Warwick (KW) and a human; the other between Warwick and a computer. Can you tell which is which?
Conversation one:
Subject: Hello. My name is Nick and I am a human.
KW: Are you happy being a human?
Subject: Yes, I am. Are you? Good. Then we are both happy.
KW: Yes, I am. Are you happy with your life at present?
Subject: Parts of it. I hate my job, but most people do, right?
KW: I'm OK with mine. What is your job?
Subject: I work as an 'online internet advertising monitor', which is fancy language for electronic filing. What do you do?
KW: I interrogate humans and machines.
Subject: Which ones do you prefer, humans or machines?
KW: Which do you prefer?
Subject: Hmm. Depends on for what purpose you mean.
KW: To go to a restaurant, for example?
Subject: Then I would much prefer going with a human.
Conversation two:
Subject: Hello, my name is Mike and I am a man.
KW: Are you happy being a human?
Subject: Judge, I'm a guy.
KW: Does that worry you?
Subject: Don't worry, we'll work everything through.
KW: Do you have worries then?
Subject: Let's move on from that basis.
KW: Why?
Subject: Why? I like this subject.
KW: What is the subject, do you think?
Subject: What is the subject I wonder if I think? I know that it's good to thonk about. Think deeplt about mysterious subjects.
KW: What mysterious subjects?
Subject: Think deeply about mysterious dubjects.
KW: What's a dubject?
Answers:
Conversation one is with a human; conversation two is with the program Ultra Hal.
No shit, Sherlock? The second conversation stops making sense in the first answer.
1. In the Turing test, as it was proposed by Turing, basically there is no way for a human to fail it. The test involves a double blind test, where each user interacts with a human and with a machine. Then if the users can't tell which of them is the human, the machine has "won". If the users correctly voted on which of them is the machine, the machine has "lost." There is no scenario there in which the human didn't pass the test. The human is the control point there, not the one taking the test.
2. But maybe you mean a test with only one entity, where basically you just have to say if that entity is too dumb to be a human.
I wouldn't really pray for that to be reasons for "disconnecting" someone. There was a story on /. a while back, titled, basically, "how I failed the Turing test."
Basically someone's ICQ number had landed on a list of sex bots. For some people that was definitive and refutable proof that he is a bot, and nothing he could say would change that. When he got one or two to ask stuff to see if he's a human, the questions were stuff where really the only correct answer for a normal human (as opposed to, say, a nerd who has to sound like he knows everything) was "I don't know." That "I don't know" was further taken as confirmation that he is a bot after all.
So do you want those people to be the ones who judge whether you live or die?
Furthermore, for most people, gullibility is akin to a deadly sin, and being fooled by a machine is akin to an admission of being terminally gullible. By comparison voting that a living human is probably a machine, counts just as being skeptical, which is actually something positive. So all things being equal, the safe vote for their self-image is that you're a machine. No matter what you say. Are you sure you want to risk life and limb on that kind of a lopsided contest?
A polar bear is a cartesian bear after a coordinate transform.
We "switch off" dogs, horses etc all the time. And these are generations ahead of any AI we have.
Personally I think we should be focusing on augmenting humans instead of creating "Real" AIs.
Why? Because we are not doing a good job taking care of the billions of already existing nonhuman intelligences. So why create even more to abuse and enslave?
Just because you can have children doesn't automatically mean the time is right. Wait till we as a civilization have grown up (to be a mature civilization) then maybe it won't be such a bad idea to have "children" of our own.
Don't forget, dogs are generally happy to obey humans and do not resent us - but this took many generations of breeding.
If we create very intelligent AIs without all the other "goodies" the "I'm so happy to see you" doggies have "built-in", we're just creating more problems rather than solutions.
In contrast if we use that sort of tech to augment humans so that they can do things better and more easily we avoid a whole bunch of potential issues.
The lines might get blurry at some point, but by that point we'd probably be more ready.
My daughter is 13 months old. She would not pass the Turing Test, yet is undeniably intelligent.
She recognizes my wife and I and all of our relatives, but is wary of strangers.
She learned cause and effect many months ago by observation: when you press a button, cool stuff happens. (We pick up the remote, she looks at the TV. We put a hand on a light switch, she looks at the light.)
She knows our relatives' names, and will look at them when you ask "Where's Charles?" or "Where's Lindsey?"
She responds to simple requests like, "Could you bring me the toy?"
She's learned how to crawl. She's learned how to walk. She's learned simple sign language for "light," "dog," "food," and "more."
I'm a long time amatuer AI hacker/researcher. I've learned more about artificial intelligence from watching my daughter develop than from my MS in CS and the bits of PhD work I did. There's an entire childhood, a virtual lifetime, of development and ability behind "carrying on a conversation." Creating a facade that does so, no matter how complex, (and we haven't even done that yet) will not be intelligent. Period. And I think it's the focus on the end results, (i.e. simulated conversation) and not on the long tedious journey it takes to create a being, that's hobbled AI research for 50 years.
True AI will never be developed if we continue to focus on the result of, and not the journey to, intelligence.