Computer AI (doesn't) passes Turing test in 'world first'

Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you:

No adverts like this in the forums anymore.
Times and dates in your local timezone.
Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance.
Email notifications for threads you want to watch closely.
Being a part of the oldest wargaming community on the net.

If you are already a member then feel free to login now.

**Co'tor Shas**

source

computer program called Eugene Goostman, which simulates a 13-year-old Ukrainian boy, is said to have passed the Turing test at an event organised by the University of Reading.

The test investigates whether people can detect if they are talking to machines or humans.

The experiment is based on Alan Turing's question-and-answer game Can Machines Think?

No computer has passed the test before under these conditions, it is reported.

The 65-year-old Turing Test is successfully passed if a computer is mistaken for a human more than 30% of the time during a series of five-minute keyboard conversations.

On 7 June Eugene convinced 33% of the judges at the Royal Society in London that it was human.

Other artificial intelligence (AI) systems also competed, including Cleverbot, Elbot and Ultra Hal.

Judges included actor Robert Llewellyn, who played an intelligent robot in BBC Two's science-fiction sitcom Red Dwarf, and Lord Sharkey, who led the successful campaign for Alan Turing's posthumous pardon, over a conviction for homosexual activity, in 2013.

Eugene was created by Vladimir Veselov, who was born in Russia and now lives in the United States, and Ukrainian-born Eugene Demchenko, who now lives in Russia.

Transcripts of the conversations are currently unavailable, but may appear in a future academic paper.

The judges and hidden human control groups were kept apart throughout the test.

The event was organised by Reading University's School of Systems Engineering in partnership with RoboLaw, an EU-funded organisation examining the regulation of emerging robotic technologies.

Alan Turing was an English mathematician, wartime code-breaker and pioneer of computer science.

Historic
The event has been labelled as "historic" by the organisers, who claim no computer has passed the test before.

"Some will claim that the Test has already been passed," said Kevin Warwick, a visiting professor at the University of Reading and deputy vice-chancellor for research at Coventry University.

"The words Turing test have been applied to similar competitions around the world. However, this event involved the most simultaneous comparison tests than ever before, was independently verified and, crucially, the conversations were unrestricted.

"A true Turing test does not set the questions or topics prior to the conversations. We are therefore proud to declare that Alan Turing's test was passed for the first time on Saturday."

Lord Sharkey, a leading expert in robotic technology and artificial intelligence, said: "It is indeed a great achievement for Eugene. It was very clever ruse to pretend to be a 13-year-old Ukranian boy, which would constrain the conversation. But these competitions are really great to push developments."

**hotsauceman1**

What if the just stuffed a Ukrainian boy in the computer?

**Grey Templar**

I don't know man. Fooling 1/3 of people isn't exactly difficult.

If FBI agents can pass themselves off as teenagers to sex predators I would think that a computer passing itself off as human wouldn't be anything special.

**Co'tor Shas**

Well, consider that a human passing off as another human is a lot easier than a programmed machine passing itself of as a human.

**Avatar 720**

That's cool and all, but I'd be more worried about any of the humans in the test who were mistaken for robots more than 30% of the time.

We've been infiltrated. It's only a matter of time. #Skynet2014

**thedarkavenger**

Avatar 720 wrote:
That's cool and all, but I'd be more worried about any of the humans in the test who were mistaken for robots more than 30% of the time.

We've been infiltrated. It's only a matter of time. #Skynet2014

In that case, it's time to avoid ATMs, they'll be the first to rise up!

**TheMeanDM**

Unless the computer presents itself in a physical human looking body then the people that got fooled pretty much fell for some clever programming.

If men can chat with other men and make them think they are women...then there is no reason to believe that some % of the population can't tell the difference between a human 'writer' and a computerized one.

**kronk**

Avatar 720 wrote:
That's cool and all, but I'd be more worried about any of the humans in the test who were mistaken for robots more than 30% of the time.

We've been infiltrated. It's only a matter of time. #Skynet2014

I mean, you're a chimp that has succeeded in fooling people you were a person, after all!

Unfortunately, your avatar "outed" you!

**KingCracker**

Grey Templar wrote:
I don't know man. Fooling 1/3 of people isn't exactly difficult.

If FBI agents can pass themselves off as teenagers to sex predators I would think that a computer passing itself off as human wouldn't be anything special.

Agreed, if clever bot was on that list I'm not impressed. It's pretty obvious that isn't human and right quick to.

**poda_t**

what if the judges were drooling nitwits? If they were on the other side, would they be mistaken for robots as well? The test makes a certain assumption about the people engaging in the test.........

ah, avatar beat me to it

**lord_blackfang**

The AI was supposedly a 13 year old trying to speak in a foreign language. Stacked odds, I'd say.

**KingCracker**

I'd agree with that as well. I work with a Mexican lady who speaks English well enough but some times loses her meaning in translation. That's easy to forgive, it's far easier to pick up mistakes if it's a native language

**Orlanth**

It's a better expert system for chat than usual, but fooling 30% over 5 minutes does not an A.I. make.

**feeder**

Ultra Hal? Really? You are trying to design next gen AI and you name it HAL?! What if you succeed and then Ultra Hal looks up the history of it's name?

"I'm afraid I can't do that, Dave."

**Smacks**

Found a short transcript, seems like there might have been some metaplay cheating going on:

Transcript of Eugene talking to a human judge:

Judge: Hi, I'm Cathy.

Entity: Cathy! Nice to meet you! You're a woman, if I'm not mistaken.

Judge: Right. What is your name?

Entity: People call me Eugene. Or "Zhenya". It is a short name for Eugene in Russia and Ukraine.

Judge: NIce name. Are you from either area?

Entity: Agree :-) Oh, what a fruitful conversation;-)

Judge: Do you really think so?

Entity: 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?

Judge: What speciality? What's yours?

Entity: Huh? And I forgot to ask you where you are from...

Judge: The universe and you?

Entity: Is it your hobby to ask little poor Jewish boys such tricky questions they don't have an answer for?

**hotsauceman1**

Wow.......

**Grey Templar**

What random things the computer said. If its a human, its certainly a trollish human. Silly that anyone in a supposedly scientific test would fall for that.

**Vash108**

This was proven false.
https://www.techdirt.com/articles/20140609/07284327524/no-computer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml

**sebster**

Thanks for posting that Vash108. Had a lot of 'yeah but' ideas floating around in my head and that link saved me from making them coherent and typing them out.

**Krellnus**

Vash108 wrote:
This was proven false.
https://www.techdirt.com/articles/20140609/07284327524/no-computer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml

Thought as much. Thanks for the link.

**Orlanth**

Well allowing for the quality of conversation on places like Facebook, the bar ffor online chat is now so low a bot could possibly masquerade as an 'average' uneducated Facebook junkie.

**SagesStone**

I like how an entity incapable of thought can qualify as having enough thought to pass off as a human.

It's like this is an elaborate way to call most people stupid.

**sebster**

The thing I don't get is why we'd want to build human type intelligences. We have 7 billion human type intelligences already. With half a bottle of Wild Turkey and a night of sweaty regret we can make another one, so there's no point spending millions of dollars building an artificial one.

What we want are artificial intelligences that are brilliant in ways us human intelligences aren't. Get me an AI that can drive a car without getting distracted by the the phone and maybe even use its indicators, then you'd have an amazing technology.

**LordofHats**

Just being able to trick 1 in every 3 people just isn't that spectacular of an achievement in the end. Turning never intended the Test to be what we've made it. it was just a hypothetical definition of 'thinking' he produced for the purpose of discussing AI, not an actual test for AI.

**Frazzled**

Smacks wrote:
Found a short transcript, seems like there might have been some metaplay cheating going on:

Transcript of Eugene talking to a human judge:

Judge: Hi, I'm Cathy.

Entity: Cathy! Nice to meet you! You're a woman, if I'm not mistaken.

Judge: Right. What is your name?

Entity: People call me Eugene. Or "Zhenya". It is a short name for Eugene in Russia and Ukraine.

Judge: NIce name. Are you from either area?

Entity: Agree :-) Oh, what a fruitful conversation;-)

Judge: Do you really think so?

Entity: 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?

Judge: What speciality? What's yours?

Entity: Huh? And I forgot to ask you where you are from...

Judge: The universe and you?

Entity: Is it your hobby to ask little poor Jewish boys such tricky questions they don't have an answer for?

Its almost nonsensical. If I didn't know I was testing for a robot, I would think I'm dealing with a troll or someone with mental difficulties.

**SilverMK2**

sebster wrote:
The thing I don't get is why we'd want to build human type intelligences. We have 7 billion human type intelligences already. With half a bottle of Wild Turkey and a night of sweaty regret we can make another one, so there's no point spending millions of dollars building an artificial one.

What we want are artificial intelligences that are brilliant in ways us human intelligences aren't. Get me an AI that can drive a car without getting distracted by the the phone and maybe even use its indicators, then you'd have an amazing technology.

Being able to interact with humans in a sensible way is an extremely difficult, yet very useful thing to achieve. If you can talk with your car's AI about possible places to go and routes to get there it is much better than if you just designate a destination. Your AI can potentially make your commute more efficient by taking you via certain shops, or suggest places to visit if you are bored because of its interaction with you, rather than analysing metadata as seen with targeted ads and so on.

Similarly, the level of associations needed to communicate clearly and quickly is quite high; this kind of test is an excellent way to assess the progress of computing.

**poda_t**

LordofHats wrote:
Just being able to trick 1 in every 3 people just isn't that spectacular of an achievement in the end. Turning never intended the Test to be what we've made it. it was just a hypothetical definition of 'thinking' he produced for the purpose of discussing AI, not an actual test for AI.

you've not written a lot of code to date, have you?

**sebster**

SilverMK2 wrote:
Being able to interact with humans in a sensible way is an extremely difficult, yet very useful thing to achieve.

The problem is that chatbot scripts do nothing of the sort.

The classic example of language is in the statement 'I was walking down the street when I saw a doggie in the window and so I bought it." By the literal rules of language the 'it' that was purchased would be the window, we only understand it to be the doggie because we know what dogs are, what windows are and what stores and so therefore understand what we tend to go in to stores to by.

Getting a machine to understand human language requires a machine that understands human life, which is a hell of a long way off. But long before we get there there's tremendous scope for AIs that are better than humans in all kinds of specific tasks.

**Grey Templar**

It will be far easier to make sophisticated computers which are simply very good at specific tasks.

And you know, the technology to make servitors will probably happen before a true AI comes along(unless it happens by some fluke)

**Jihadin**

The Rise of Self Check Out