Turing Test

The Turing Test is a behavioural test of whether a machine can imitate human conversation. It evaluates whether an artificial intelligence system can display human-like conversational behaviour such that a human evaluator cannot reliably distinguish it from a human participant.

Origin and test procedure

The test was proposed by Alan Turing in his 1950 paper, “Computing Machinery and Intelligence”, through the idea of an imitation game.

• A human evaluator conducts text-only conversations with a hidden machine and a hidden human, avoiding visual and auditory clues.
• The evaluator attempts to identify which participant is the machine based solely on their responses.
• A machine is considered successful when its conversational behaviour cannot be reliably distinguished from that of the human participant.

Capability being evaluated

The test operationalises intelligence as observable conversational behaviour, rather than examining how the machine is designed internally.

• Success can require natural-language processing, reasoning, knowledge representation and learning.
• The central capability is producing contextually appropriate, human-like responses, not merely supplying predetermined answers.

What success does not establish

Passing the test shows an ability to simulate human conversation under the test conditions. It does not by itself prove consciousness, emotions, genuine understanding, factual reliability or general intelligence.