What is the Turing Test that determines if computers can think?

The game involves a human guessing if a player is a computer or another human.

July 21, 2023, 6:10 AM

The Turing Test is arguably one of the most well-known methods of evaluating how well artificial intelligence (AI) can think like a human.

The test was developed by British mathematician and computer scientist Alan Turing, who had been studying the concept of machine learning for several years beforehand.

Turing believed that a computer could be described as intelligent if it can mimic human responses under specific conditions.

"The thing to remember is computers were giant room-based, vacuum-tube devices," Mark Reidl, associate director at the Georgia Tech Machine Learning Center, told ABC News. "But they were doing things that humans were doing, particularly breaking codes, solving math problems, things like that."

He continued, "So they started to think, 'Can this go further than that? What can they do that humans can do? Can they do what humans can do?'"

What is the Turing Test?

The Turing Test involves three players: a computer, a human respondent and a human interrogator. All three are placed in separate rooms or in the same room but physically separated by terminals.

The interrogator asks both players a series of questions and, after a period, tries to determine which player is the human and which is the computer.

Turing Test
Turing Test
ABC News Photo Illustration, Getty Images

If the interrogator fails to determine which player is which, the computer is declared the winner and the machine is described as being able to think.

History of the Turing Test

"'I propose to consider the question, 'Can machines think?'" Turing opened his 1950 paper, "Computing Machinery and Intelligence."

It was in this paper that he laid out the thought experiment for the test, but which he called, "The Imitation Game," based on a party game with the same name.

In the original game, a man and a woman go into separate rooms from the rest of the party. Guests feed them a series of questions and try to determine from the written or typewritten answers which is the man and the woman.

One of the players is trying to trick the interrogator and the other player is trying to help the interrogator.

For Turing's new version of the game, he described it as, "We now ask the question, 'What will happen when a machine takes the part of A in this game?' Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, 'Can machines think?'"

Turing Test today

Despite being developed more than 70 years ago, the Turing Test Is still used today to assess artificial intelligence.

In 1966, a German programmer created ELIZA -- a program now considered a chatbot -- which replicated the behavior of a psychologist and was considered the first to potentially have passed the Turing Test.

More recently, in 2014, a computer program named Eugene Goodman, which simulates a teenage Ukrainian boy, was found to pass the Turing test as did Google's LaMDA in 2022.

PHOTO: Alan Turing (1912-1954), computer scientist and cryptologist instrumental in breaking Germany's 'enigma' machine code during World War II, c. 1928.
Alan Turing (1912-1954), computer scientist and cryptologist instrumental in breaking Germany's 'enigma' machine code during World War II, c. 1928. Alan Mathison Turing was a British pioneering computer scientist, mathematician, logician, cryptanalyst, philosopher, mathematical biologist, and marathon and ultra distance runner. He was highly influential in the development of computer science.
Universal Images Group via Getty Images

Experts say the test isn't perfect, however. The Turing test is "superficial," Reidl says, and even though more powerful technologies are being built, much of artificial intelligence can't do what humans can do.

"We have reasonably good self- driving cars, we have chess programs, we have programs that can play computer games…but they're not the same artificial intelligence," he said. "They're not general. They can't do more than one thing."

Reidl added, "We're starting to see some broadness, although broadness may be in language. I still don't think we see things driving cars and talking and telling poetry. But we start to see some of the hints that we're starting ot understand how to build broader technologies."