CHINA TOPIX

12/22/2024 10:06:22 pm

Make CT Your Homepage

Artificial Intelligence Passes Turing Test, Fools Judges Who Thought It's a 13-Year-Old Boy

Last Saturday, June 7, a historic breakthrough occurred at the Royal Society of London when an artificial intelligence passed the Turing Test for the first time and fooled a third of the judges that it was a human being. 

According to the UK's University of Reading, a computer program named Eugene Goostman was able to convince a third of the judges that it was a 13-year-old boy. 

Like Us on Facebook

The milestone was the first of its kind and the supercomputer the first to have passed the Turing Test. The test is an experiment designed to find out whether people can determine if they are communicating with a person or a machine. The experiment uses text as a means of conversation. According to the experiment, the computer would have passed the test if it succeeds in convincing humans it is one of them. 

The test was proposed by Alan Turing during the 1950s. Turing is regarded by the computing community as the father of artificial intelligence (AI). 

Vladimir Veselov and Eugene Demchenko developed Eugene and had the program competed against four other supercomputers at the event. Veselov and Demchenko designed Eugene to mimic the responses of a 13-year-old boy. 

"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. 

Veselov and Demchenko plans to make Eugene much smarter. The two said they will continue working on how they can improve Eugene's conversation logic. 

According to university organizers, Eugene was successful at convincing 33 percent, or a third, of all human judges present at the Royal Society event. The university said the results of Eugene's test were independently verified. 

University of Reading's visiting professor of cybernetics Kevin Warwick said in a statement that the Royal Society event did not put so many restrictions on topics and questions that can be asked. He said the open-endedness of the questions and topics is the true measure of the Turing Test. 

In the 1950 paper titled 'Computing Machinery and Intelligence,' Alan Turing raised the question 'can machines think?' He also introduced the concept of allowing humans to guess whether they're having conversations with a machine or a person using only text messages.

Real Time Analytics