Turing Test

Free AI web copilot to create summaries, insights and extended knowledge, download it at here

1084

Abstract

The elementary Turing test uses three terminals. Two are controlled by people and the third by a computer. Each terminal is physically separate from the other two. One person is the interrogator, while the other person and the computer answer them.The interrogator asks questions to the person and the computer in a predetermined form, in a given domain and context, and for a predefined period of time, for example, 10 minutes.At the end of the allotted time, he designates which terminal is, in his opinion controlled by a person and which by a computer.The test is repeated many times. If the interrogator is correct in half of the test runs or less, the computer is deemed to have artificial intelligence, because its interrogator considers it human as a person.However, the Turing test is criticised, in particular, because the nature of the questions is necessarily limited so that the computer can present the traits of an intelligence resembling ours.For example, a computer can get a high s

Options

core if the questions fall within a narrow area of knowledge, such as number theory in mathematics, and expect a yes or no answer.However, in response to broader questions, such as in a conversation, a computer would be less likely to respond as a person does. This observation is particularly true if it is a subject with a strong emotional, or socially sensitive load. In certain specific cases, if the machine's performance greatly exceeds in quality and speed that of a person, the interrogator will easily know how to decide between them.For example, Google’s search engine would significantly outperform a person in a Turing test targeting information searches. For many researchers, the question of whether or not a computer can pass a Turing test is no longer relevant. Instead of focusing on convincing someone that they are conversing with a human and not with a computer program, the focus should be on how to make human-machine interaction more intuitive and efficient- using a conversational interface, for example.</article></body>

Brief Intro

In artificial intelligence ( AI ), the Turing test is a method of determining whether a computer is capable of thinking like a human.

Its name refers to Alan Turing, an English mathematician who pioneered artificial intelligence in the 1940s and 1950s and owed the first version of the test.

According to such tests, a computer is gifted with artificial intelligence if it can mimic a human’s responses under given circumstances. In the Turing test, if the person doing the test cannot consistently determine whether it is a computer or a responding person, the computer is deemed to have “passed” the test.

The interrogator asks questions to the person and the computer in a predetermined form, in a given domain and context, and for a predefined period of time, for example, 10 minutes.

At the end of the allotted time, he designates which terminal is, in his opinion controlled by a person and which by a computer.

The test is repeated many times. If the interrogator is correct in half of the test runs or less, the computer is deemed to have artificial intelligence, because its interrogator considers it human as a person.

However, the Turing test is criticised, in particular, because the nature of the questions is necessarily limited so that the computer can present the traits of an intelligence resembling ours.

For example, a computer can get a high score if the questions fall within a narrow area of knowledge, such as number theory in mathematics, and expect a yes or no answer.

However, in response to broader questions, such as in a conversation, a computer would be less likely to respond as a person does. This observation is particularly true if it is a subject with a strong emotional, or socially sensitive load. In certain specific cases, if the machine's performance greatly exceeds in quality and speed that of a person, the interrogator will easily know how to decide between them.

For example, Google’s search engine would significantly outperform a person in a Turing test targeting information searches. For many researchers, the question of whether or not a computer can pass a Turing test is no longer relevant. Instead of focusing on convincing someone that they are conversing with a human and not with a computer program, the focus should be on how to make human-machine interaction more intuitive and efficient- using a conversational interface, for example.