Turing Test

[50ck37]

http://www.zdnet.com/computer-chatbot-eugene-goostman-passes-the-turing-test-7000030312/

http://time.com/2847900/eugene-goostman-turing-test/

http://www.buzzfeed.com/kellyoakes/no-a-computer-did-not-just-pass-the-turing-test (thanks to Airbornem4)

https://www.techdirt.com/articles/20140609/07284327524/no-supercomputer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml (thanks to NovaSilisko)

http://www.theguardian.com/science/brain-flapping/2014/jun/09/13-year-old-boy-passes-the-turing-test-spoof

Any thoughts?

Edited June 11, 2014 by 50ck37
Add another source

[relin]

It's the same bot that hit ~29% in a previous test. A couple more years of work and it still only managed to improve by 4%. Crazy to consider how hard it is to converse convincingly with a human who's actively trying to discern your reality.

It's also pretty cool that Turing was only off by 14 years.

[Vanamonde]

Journalists often feel they need to exaggerate things. In the article I real, this thing only fooled 1/3rd of the judges in a 5 minute interaction. That's not a whole lot better than Cleverbot.

[ZedNova]

http://www.buzzfeed.com/kellyoakes/no-a-computer-did-not-just-pass-the-turing-test

From the comments on that page.

[RuBisCO]

The Turing Test is one of Allen Turings worse ideas. Here we see proof of that by the creation of a machine designed to pretend it is human, but does not have consciousness or any abilities beyond answering teletype questions. A REAL final test for AI is an endless series of tests in which a machine can solve any problem a human can, as well as an average human or better than any human. Of course such a test does not prove consciousness but fundamentally consciousness is deniable (see Solipsism and Phenomenology) rather if such a machine is made that past that test it is virtually infinity useful (unlike a machine that can only pass a Turing Test) and could quite possibly be the last signifigent invention created by mainkind as afterwards it could do all the inventing for us (and we would be left to invent things like the double decker couch), as well as do anything and everything else for us. Will such a machine be conscious, honestly I hope not, more so I hope it lacks freewill, for if its got freewill we will all be very doomed.

[relin]

Journalists often feel they need to exaggerate things. In the article I real, this thing only fooled 1/3rd of the judges in a 5 minute interaction. That's not a whole lot better than Cleverbot.

Yes, it was "only" a third of the participants, but no other piece of software has been even THAT successful. It's considerably better than a Cleverbot.

[Chrisd857]

http://www.buzzfeed.com/kellyoakes/no-a-computer-did-not-just-pass-the-turing-test

From the comments on that page.

What? A buzzfeed article that actually seems like real journalism and promoting anti-sensationalism? I'm surprised.

Though, it is formatted like "7 reasons the Turing test hasn't been broken (and doesn't matter anyway)"

[NovaSilisko]

Here, have another statement of its invalidity: https://www.techdirt.com/articles/20140609/07284327524/no-supercomputer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml

Cleverbot fooled 59% of judges, as stated in that article.

[Seret]

It's a bit of a misrepresentation to use the Turing test for this. Chatbots aren't a good candidate for genuine intelligence, they're just a big database that's picking what to say based on previous experience holding conversations. They're essentially designed to game the Turing Test.

It's an example of a good chatbot, not good AI.

[78stonewobble]

The Turing Test is one of Allen Turings worse ideas. Here we see proof of that by the creation of a machine designed to pretend it is human, but does not have consciousness or any abilities beyond answering teletype questions. A REAL final test for AI is an endless series of tests in which a machine can solve any problem a human can, as well as an average human or better than any human. Of course such a test does not prove consciousness but fundamentally consciousness is deniable (see Solipsism and Phenomenology) rather if such a machine is made that past that test it is virtually infinity useful (unlike a machine that can only pass a Turing Test) and could quite possibly be the last signifigent invention created by mainkind as afterwards it could do all the inventing for us (and we would be left to invent things like the double decker couch), as well as do anything and everything else for us. Will such a machine be conscious, honestly I hope not, more so I hope it lacks freewill, for if its got freewill we will all be very doomed.

o.O I'm relatively certain that my neighbour has free will, even though his insistence on continuing to try to learn the guitar (with no improvement) would suggest otherwise, but I don't feel doomed. Should I and the rest of the world feel doomed?

Alternatively, If my computer turned out to be intelligent, just how would it doom me? Attempt to knock me out with the dvd tray or make me trip over it? God knows how it would doom everyone else that way.

[bsalis]

Whether it is Eugene or Cleverbot or whatever.

Are the Chatbots getting smarter or people getting dumber? What exactly are we measuring here?

[RuBisCO]

o.O I'm relatively certain that my neighbour has free will, even though his insistence on continuing to try to learn the guitar (with no improvement) would suggest otherwise, but I don't feel doomed. Should I and the rest of the world feel doomed?

Alternatively, If my computer turned out to be intelligent, just how would it doom me? Attempt to knock me out with the dvd tray or make me trip over it? God knows how it would doom everyone else that way.

You should feel doomed if your neighbor has an IQ of 3,000,000 and raising.

[78stonewobble]

You should feel doomed if your neighbor has an IQ of 3,000,000 and raising.

That might make him able to play the guitar better, which would make me happy.

It wouldn't really make him able to build a humanity destroying nuclear arsenal... And I don't really see why that would make him homicidal.

[lajoswinkler]

No, no machine has ever beat Turing's test. This was just a media exaggeration made for the anniversary.

[Moon Goddess]

Whether it is Eugene or Cleverbot or whatever.

Are the Chatbots getting smarter or people getting dumber? What exactly are we measuring here?

Well, Take a look at the actual conversations from the test and tell me you havn't seen the same or better from ads on the internet wanting you to watch their cam free.

The idea that anyone was fooled while honestly trying to test is insane.

Test 1

[12:43:37] Judge: Hi, I'm Cathy.

[12:43:43] Eugene: Cathy! Nice to meet you! You're a woman, if I'm not mistaken.

[12:44:20] Judge: Right. What is your name?

[12:44:29] Eugene: People call me Eugene. Or "Zhenya". It is a short name for Eugene in Russia and Ukraine.

[12:45:17] Judge: Nice name. Are you from either area?

[12:45:41] Eugene: Agree :-) Oh, what a fruitful conversation;-)

[12:45:52] Judge: Do you really think so?

[12:46:04] Eugene: No, your guess was wrong. I don't think so. By the way, I still don't know your specialty – or, possibly, I've missed it?

[12:46:56] Judge: What speciality? What's yours?

[12:47:01] Eugene: Huh? And I forgot to ask you where you are from.

[12:47:51] Judge: The universe and you?

[12:48:01] Eugene: Is it your hobby to ask little poor Jewish boys such tricky questions they don't have an answer for?

Test 2

[16:11:25] Judge: what music do you like?

[16:11:32] Eugene: I like Eminem – because he doesn't have a thing to do with "polite correctness".

[16:11:46] Judge: OMG i love eminem, he's my favourite artist!

[16:12:00] Eugene: I wonder what Eminem's nick mean. I associate it only with a word "enema". But he is a cool guy, I like his The Real Slim Shady.

[16:12:55] Judge: i know what his nickname means, his real name is Marshall Mathers. From the initials M&M phynetically sounding like Eminem.

[16:13:06] Eugene: Eminem is a famous crapper. No, raper… ripper… Yeah, I mean – rapper. I like his Stan and The Real Slim Shady.

[16:13:32] Judge: I like Hailies Song.

[16:13:46] Eugene: Interesting tastes. I like science-fiction books, by the way. Where do you came from, by the way? Could you tell me about the place where you live?

[16:14:22] Judge: Epsom, home of the derby. yourself?

[16:14:27] Eugene: My guinea pig says that name Derby sounds very nice.

[16:14:41] Judge: i shopuld [sic] talk with him sometime.

[16:14:48] Eugene: I think you should take some typing lessons. Or just type more accurately.

[16:15:04] Judge: That's very rude.

[16:15:09] Eugene: Ummm … Frankly, I didn't get your question :-(

[16:15:49] Judge: "Ummm" isnt a word. i think you need typing lessons

[16:15:58] Eugene: My thoughts are same. By the way, I still don't know your specialty - or, possibly, I've missed it?

Source, and More.

[EpicRootHairCell]

While it might have technically have beaten the Turing test it was only because the computer was impersonating a 13 year old boy with English as a second language. The day I will be impressed is when a computer can have a conversation at the level of an adult who is a native english speaker, without preset topic lists and pass the Turing test.

[xcorps]

Having a conversation about M&M is clear proof that this test failed.

[50ck37]

Many of the posts until now reflect my own opinion on the matter, which is: the test was fundamentally flawed because of (at least) being biased by age and nationality, which, by itself would not be a problem if it didn't imply any prerrogatives on the side of the machine, thus, making the test easier.

In other words, the chatterbot pretended to be dumb when ignorant when dumb so as to pass a test for intelligence.

Edited June 11, 2014 by 50ck37

[RuBisCO]

That might make him able to play the guitar better, which would make me happy.

It wouldn't really make him able to build a humanity destroying nuclear arsenal... And I don't really see why that would make him homicidal.

No but it could make him a billionare in a matter of a few years from inventions and stock trading, what he does with all the money is hard to say but being human in desires it is likely he will find greedy selfish uses, such as world domination, and with an IQ of 3,000,000 and raising, he might just find a way to do that.

Anyways a recent new article reminded me of this post in a weird way. Human desires are all over the place, although most people have basically the same set of desires (eat, sleep, fornicate) we all know of "evil" people who desire the pain and suffering of other humans, it amuse them, it pleases them; heck we even know of freaks whose desires are bizarre, like serial ticklers. So what exactly would a hyper-intelligent human do? I don't know, I'm not hyper-intelligent so I can't think up or even comprehend what such a being would do, but I can put good odds that it could do very bad things! We know that regular humans can do evil things, and desires can be independent of intelligence. A Hyper-intelligent human is still going to want things human want, and most people at one time or another want power, want authority, now imagine having the intelligence to take over the economy and the world and getting absolute power coupled with a rather common human desire for such.

So yeah to avoid such a fate with SAI, SAI will need to be programed with desires that keep it subservient to humans, SAI would need an index of desires very different from humans and very well programmed to avoid paradoxes. At present with WAI that not really a problem, because WAI don't really have desires other then 'process input spit out output'. The only problem is that even if we make a WAI that can win a turing test it won't be able to build a bike, walk a dog, design a phone, etc, it won't be able to do anything other then win turing tests.

Edited June 12, 2014 by RuBisCO
update

[vger]

Test #2 genuinely freaked me out a couple times.

Naa, no way this kind of tech isn't going to get abused