Hello, today I will talk about the Turing test, created by Alan Turing, considered the father of computer science, to answer the question "Can machines think?"
You can click on the following topics to jump to the section where I explain them:
What is the Turing Test?
Can machines really think? And if they can, how can we prove it? In 1950, Alan Turing introduced the Turing test in his paper "Computing Machinery and Intelligence" to logically answer these questions. Although the test has been widely criticized over the years, it still holds an important place in the philosophy of artificial intelligence. So, how is the Turing test conducted?
In the test, there is initially a human evaluator. This evaluator is asked to interact with different individuals through a chat application. The interaction must be in written form, not through voice or video. The group the evaluator interacts with can either consist of only humans or a mix of humans and a machine undergoing the Turing test. The evaluator engages in written conversations with the group for 5 minutes. If the evaluator can identify a machine in the group, the machine fails the test. If there is a machine in the group and the evaluator believes that everyone they interacted with is human, the machine successfully passes the Turing test.
Sixteen years after the test was published, in 1966, Joseph Weizenbaum created a program called Eliza that succeeded in passing the test on some individuals. Eliza essentially behaved like a psychotherapist. It would automatically respond with pre-prepared sentences if it found previously learned keywords in the user's sentences. If no such keywords were found, it used stock responses. Fundamentally, Eliza aimed to prompt the user to talk more by asking questions. This approach convinced many people that Eliza was a real psychotherapist.
Examples of the Turing Test
In this section, you can also read the dialogues and try to guess which one is the machine and which one is the human. I will show three examples in total. In these tests, the judges are initially informed that one is a human and the other is a machine, and they are expected to identify accordingly.
Jury's Guess: The jury guessed that both conversations were with real humans.
Actual Result:
The conversation on the left is with an adult male, and the conversation on the right is with a machine named Eugene Goostman.
Jury's Guess:
The jury said that the conversation on the left was with a young male and the one on the right was definitely with a machine.
Actual Result:
The conversation on the left is with a machine named Eugene Goostman, and the conversation on the right is with a real human.
Jury's Guess:
The jury guessed that the conversation on the left was with a machine and was unsure about the one on the right.
Actual Result:
The conversation on the left is with a machine named Ultra Hal, and the conversation on the right is with a real human.
Source: