A supercomputer has achieved an artificial intelligence milestone by passing the Turing Test, according to the University of Reading in the U.K.
At an event on Saturday at the Royal Society in London, a conversation program running on a computer called Eugene Goostman was able to convince more than a third of the judges that it was human.
It marks the first time that any machine has passed the Turing Test proposed in 1950 by Alan Turing, regarded as the father of artificial intelligence (AI), according to the university, which organized the event.
The Turing Test is an experiment that focuses on whether people can tell whether they are communicating with a person or a machine. If the machine is able to fool people into thinking it’s human during a series of text conversations, it’s considered to have passed the test.
The program, dubbed Eugene, was developed by Vladimir Veselov and Eugene Demchenko and competed against four other supercomputers at the Royal Society event. Eugene is designed to simulate the responses of a 13-year-old boy.
Eugene convinced 33 percent of the human judges that it was human, and the results were independently verified, the university said.
“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,” Veselov said in a statement.
“Going forward, we plan to make Eugene smarter and continue working on improving what we refer to as ‘conversation logic.’”
The event follows many AI competitions such as the Loebner Prize in which programs known as chatbots vie for the most humanlike performance.
Kevin Warwick, a visiting professor of cybernetics at the University of Reading, said in a statement that the event at the Royal Society did not place restrictions on topics or questions, and was thus a true Turing Test.
Turing’s 1950 paper “Computing Machinery and Intelligence” introduced the question “Can machines think?” as well as the concept of guessing if it is a machine or person based on text messages alone.
Various versions of the test have set different rules. The test has been controversial for seeking to find out whether machines can be intelligent in a human sense. It has also produced vigorous debate about the nature of intelligence itself.
“Since its inception, AI has been defined by milestone after milestone, be it playing chess or learning to parse sentence structure,” Columbia University computer science professor Tony Jebara wrote in an email, hailing the result.
“Time and time again, we have said ‘Once a computer could do X, we would have created intelligence.’ However, each time a milestone was achieved, we turned around and said ‘This is still not AI’ and once again raised the bar. We still have much to do and much to learn about how to make computers intelligent and how our own minds work.”