Turing test passed by computer

A computer has passed the Turing test, an artificial intelligence measure of human-like capability devised more than half a century ago, the University of Reading has announced.

'Eugene Goostman' designed to imitate 13-year-old boy

Eugene Goostman, a 'chatbot' software program designed to talk like a 13-year-old boy, became the first computer to pass the Turing test by convincing 30 per cent of its chat partners it was human. (Princeton AI)

A computer has passed the Turing test, an artificial intelligence measure of human-like capability devised more than half a century ago, the University of Reading has announced.

On Sunday, Eugene Goostman, a "chatbot" software program designed to talk like a 13-year-old boy, became the first to pass the test, the university said in a news release.

The program, created by Vladimir Veselov and Eugene Demchenko at U.S.-based Princeton Artificial Intelligence, convinced more than 30 per cent of its conversation partners and judges that it was human in an independently verified Turing test consisting of 30 simultaneous, unrestricted conversations.

The achievement is a "historic milestone in artificial intelligence," added the university, which co-organized the event in London where the test took place, along with RoboLaw, an European organization that researches the regulation of emerging robotic technologies.

The Turing test is based on a question and answer game proposed by renowned British mathematician and codebreaker Alan Turing, who died exactly 60 years ago Sunday, to distinguish humans from computers.

Turing predicted in a 1950 paper that within 50 years, computers would play the game so well that an "average interrogator will not have more than 70 per cent chance of making the right identification after five minutes of questioning."

The University of Reading set its rules for the competition accordingly. Each chat took five minutes, and a chatbot could win by convincing 30 per cent of judges that it was human.

Eugene Goostman convinced 10 of the 30 judges — 33 per cent — that it was a real boy.

The University of Reading noted that other computers have passed similar tests in the past, but none with the same rigorous rules. Five other computers competed alongside Eugene Goostman, but none of the others passed the test.

Veselov acknowledged that Eugene Goostman's supposed age was part of his team's strategy.

"Our main idea was that he can claim that he knows anything, but his age also makes it perfectly reasonable that he doesn't know everything," said Veselov in a statement. "We spent a lot of time developing a character with a believable personality."

The team had been working to develop "Eugene" since 2001.

Test standard questioned

Some critics, however, disagree that Eugene passed "the" Turing test. Murray Shanahan, a professor of cognitive robotics at Imperial College London, told the technology website Buzzfeed that he thinks the program still falls short. 

"Of course the Turing Test hasn’t been passed. I think it's a great shame it has been reported that way.… We are still a very long way from achieving human-level [artificial intelligence], and it trivializes Turing’s thought experiment (which is fraught with problems anyway) to suggest otherwise."

The debate arises because Turing originally framed the test by asking whether the questioner would be able to distinguish between a computer and a human as easily as between a man and a woman, and only brought up the 30 per cent threshold later in the paper when talking about his prediction about computers' abilities in 2000.\

'A wakeup call to cybercrime'

Nevertheless, Kevin Warwick, a visiting professor at the University of Reading who specializes in the field of artificial intelligence, said in a statement that the test has implications for society.

"Having a computer that can trick a human into thinking that someone, or even something, is a person we trust is a wakeup call to cybercrime," he said. "It is important to understand more fully how online, real-time communication of this type can influence an individual human in such a way that they are fooled into believing something is true … when in fact it is not."

Just a few years after Turing published his famous paper, on June 7, 1954, he killed himself after being convicted of gross indecency for having sex with a man and being forcibly treated with female hormones to reduce his sex drive. He received a posthumous apology from British Prime Minister Gordon Brown in 2009 and was granted a royal pardon in December.

Besides his work on artificial intelligence, Turing was known for helping break Nazi Germany's Enigma code during the Second World War, and one of the top prizes in computer science, the Turing Award, is named after him.

The Turing test is also the basis for the Loebner Prize for artificial intelligence competition that takes place annually with more stringent requirements.


To encourage thoughtful and respectful conversations, first and last names will appear with each submission to CBC/Radio-Canada's online communities (except in children and youth-oriented communities). Pseudonyms will no longer be permitted.

By submitting a comment, you accept that CBC has the right to reproduce and publish that comment in whole or in part, in any manner CBC chooses. Please note that CBC does not endorse the opinions expressed in comments. Comments on this story are moderated according to our Submission Guidelines. Comments are welcome while open. We reserve the right to close comments at any time.

Become a CBC Member

Join the conversation  Create account

Already have an account?