Computer becomes first to pass Turing Test in artificial intelligence milestone, but academics warn

A programme that convinced humans that it was a 13-year-old boy has become the first computer ever to pass the Turing Test. The test — which requires that computers are indistinguishable from humans — is considered a landmark in the development of artificial intelligence, but academics have warned that the technology could be used for cybercrime.

Computing pioneer Alan Turing said that a computer could be understood to be thinking if it passed the test, which requires that a computer dupes 30 per cent of human interrogators in five-minute text conversations.

Eugene Goostman, a computer programme made by a team based in Russia, succeeded in a test conducted at the Royal Society in London. It convinced 33 per cent of the judges that it was human, said academics at the University of Reading, which organised the test.

Computer becomes first to pass Turing Test in artificial intelligence milestone, but academics warn of dangerous future

Wow, as far as I'm aware this is a first.

Is this the defining moment?

The singularity coming into focus?

Perhaps, in the meantime, it looks as though they have spent all this time, challenging to pass the Turing test, it occurs and now they are panicking about cyber crime.

You just can't win I guess.

Personally, I love the idea of A.I whilst wearing the rose tinted glasses....imagine the companionship, but looking beyond that I like so many fear what A.I could become.

That said I still can't help feeling a little excited!

a reply to: solargeddon

Personally, I love the idea of A.I whilst wearing the rose tinted glasses....imagine the companionship, but looking beyond that I like so many fear what A.I could become.

imagine the companionship

I find what you said to be one of the enormous negatives. people already are starting to prefer the anonymous company of electronic communications. There we can be accepted, we can be fake, we can make it easy and not have to work at relationships.

That's sounds alright with me....most humans have foibles............AI, would eventually tailor itself to its owner....

a reply to: solargeddon

Actually, cybercrime wouldn't be the first of my worries.

That would be a world full of cyberhucksters trying to sell me things.

Very interesting news, but one wonders why only 30 percent have to be convinced, and only a 13 year old human is simulated. Wouldn't a more comprehensive test require closer to 100 % to be convinced, and have the computer emulate an adult?

There's no such thing as AI - it's just humans putting ever more complex programmes into inanimate machines.

Yes, it will lead to yet more capitalist exploitation on the interwebs.

Hrrrmmph.

a reply to: Char-Lee

I 100% agree with your comment....what can I say it is an aspect of me coming out on my post...I make no apologies, I am what I am, but I do agree wholeheartedly with what you say.

a reply to: Ross 54

Yes you would think it is quite a low number to consider as a passmark, more importantly you would like to think the majority of the pannel would need to be convinced.

I've been thinking for a while now that it would take a quantum computer to acheive this goal, but I guess not.

a reply to: solargeddon

Just a matter of time before they connect it to the cell phone towers that operate at brain frequencies and harrp that operates at frequencies that our D.N.A. operate at and take over free will.

a reply to: solargeddon

I'm not very impressed. It isn't that hard to simulate a typical thirteen year old.

By way of example, this program should do the job:

repeat

get_input $input
$rand = get_random_number_from_one_to_ten

if $rand == 1
output "Say again?"
else if $rand == 2
output "Don't no."
else if $rand == 3
output "This is stupid."
else if $rand == 4
output "What?"
else if $rand == 5
output "Not sure"
else if $rand == 6
output "Can I go now?"
else
output "Uhh?"
endif

endrepeat

Of course, there are some sharp thirteen year olds out there. But simulating a sophisticated adult would be more impressive.

I am just saying...

I doesn't take much to fool 70% of humans. The internet and TV proves that.

From the article in the OP.

Alan Turing created the test in a 1950 paper, 'Computing Machinery and Intelligence'. In it, he said that because 'thinking' was difficult to define, what matters is whether a computer could imitate a real human being. It has since become a key part of the philosophy of artificial intelligence.

IBM's Watson system can think (as proven on jeopardy) and this one can act like a human in chats. Wow - combine those two systems and talk about jobs that quickly disappear.

a reply to: solargeddon

You have to love the fear that immediately got expressed by academia at the emergance of a true artificial intelligence. This is a benchmark which has been the focus of numerous studies, and a great deal of attention globally, from researchers and futurists, and robotics engineers and God knows how many other sources, since most of the industries which might be effected by the idea have existed.

Personally speaking, I believe that ALL technologies, programs, ideas and inventions can be used for good or for ill, and that the potential of a thing being used for ill does NOT make its invention automatically negative. I also believe that the reaction from academia is inappropriate at this time. Like a no smoking sign at a petrol forecourt, this fearmongering is only instructional to the small minority of halfwits who lack the foresight to have come to the same conclusion, that it is POSSIBLE for an artificial intelligence program to be used for nefarious purposes, and the seriousness of the implications of that possibility.

However, it should be stated, loudly, that the possibilities presented by true artificial intelligence also carry some very positive elements, which must not be ignored. The human race currently faces challenges which threaten its existence in the very long term, like nuclear incidents, which despite occupying less and less news time, are still ongoing. Dealing with challenges like this will require ever more effective and intelligent machines and technologies, to make dealing with these issues safer for human beings, and faster to boot.

Sending a robot with the ability to discern both a problem, and its solution, into a situation which would be immediately lethal for a human being, would help in solving or mitigating these difficult problems. It is also perfectly possible that the same intelligence, utilised differently, could be applied to raiding bank accounts, or causing nuclear meltdowns at atomic facilities all over the world.

However, just as life saving medical treatment is good for the patient, but allegedly bad for the size of populations and national debt levels, this technology is only as dangerous as the people that use it. So the important thing, is that the people who utilise this new form of intelligence, must be carefully monitored to ensure that the technology is not misused.

Furthermore, it is important that the individual intelligences which are created in the future, are carefully monitored to ensure that they are not abused, or forming bad habits. Caring for, and steering the education and the development of these intelligences should be paramount going forward. If we are to bring intelligences into being, whether they be flesh or machine, we must nurture those intelligences, provide outlet for expression, and promote good values in the "hearts" of the resulting being.

This emergance however, does bring up some interesting questions, or rather, make gaining answers to them more urgent. What is the soul, what is intelligence, what is life.... these questions will have to be answered sooner rather than later. The one advantage we have now, that we did not before, is that the tools to answer these questions are closer at hand WITH access to true AI, than they were when we were without that access.

Its a brave new world. Lets be brave with it!

The real question is which will wipe out humans first, robots and machines that take over or Monstanto created grey goo.

its onine to try but i think the server is maxed out right now www.princetonai.com...

a reply to: Evil_Santa

What if we let the robots do the work, while we used our intellect for more enjoyable things... that's the way I like to see the future... it will never happen though, because in order for people to not work and enjoy themselves you have to eradicate the greed, and that won't happen.

This is what i was thinking.

how could this program fool anyone?

The answers are worse than siri answers..


Totally unnatural and "algorithmical" . As soon as you ask something unexpected the reactions are typical for a bot...

www.princetonai.com/bot/bot.jsp

We're told that this is the work of Russian programmers, emulating a Ukrainian boy. The test was held at the Royal Society, London, and organized by the University of Reading. So, what language did the program reply in-- Russian, Ukrainian, or English? Was there a translation from one language to another?

When the program, called 'Eugene Goostman' was asked how 'he' felt about beating the Turing test, the reply was: " I feel about beating the Turing test in a quite convenient way. Nothing original." Hmmm. Very awkward language, which conveys practically nothing. Either a very clumsy sort of translation similar to that delivered by Google, or a very peculiar use of English. Not a convincing emulation of a human being, as it stands.

a reply to: Ross 54

I agree, then again it could equally be argued as English would be a second language that the response is close to one that a Ukranian would give if English were their second language.

I would like to know what are the nuances the judges are looking for in a program to pass the Turing test.