---
title: 'Turing Test'
description: 'An imitation game in which a human judge uses text conversation to assess whether a machine can be distinguished from a human participant.'
canonical_url: 'https://darkfactory.dev/glossary/turing-test'
markdown_url: 'https://darkfactory.dev/glossary/turing-test.md'
collection: glossary
date_published: '2026-08-04T00:00:00-04:00'
date_modified: '2026-08-26T00:00:00-04:00'
---

# Turing Test


## Definition

The Turing Test is the later name for an imitation game Alan Turing proposed in his 1950 paper "Computing Machinery and Intelligence." In the machine version, an interrogator communicates through text and tries to distinguish a machine from a human participant. The text channel removes physical appearance and voice from the judgment.

Turing introduced the game after arguing that the question "Can machines think?" was too ambiguous to settle through definitions. He proposed examining an observable conversational performance instead. The paper discusses learning machines, objections, and a prediction about how well machines might play the game by the end of the century. It does not supply one universal contest protocol or the simple 30-percent deception threshold sometimes attributed to it.

Turing called his setup the "imitation game." The sources reviewed do not establish who first shortened the later discussion to "Turing Test," so the glossary credits Turing with the proposal rather than claiming he coined the eponym himself.

## Distinguish it from nearby terms

- The test evaluates **human-like conversational imitation** under a particular judge, time limit, topic range, and protocol.
- It does not directly test consciousness, subjective experience, truthfulness, safety, physical competence, or performance on tasks outside the conversation.
- A **CAPTCHA** asks whether a user is human, often through an automated challenge. It reverses parts of the setting and is not the original imitation game.
- Modern claims that a system "passed" may use different rules. Results are not comparable until the protocol and judge population are specified.

## Operational significance

The test remains valuable as a thought experiment because it replaces an unobservable internal claim with a behavioral procedure. It is a poor acceptance test for most deployed systems. A medical assistant, coding agent, or financial model should be evaluated for task accuracy, calibration, security, and appropriate refusal, even if its conversation sounds entirely human.

## Check your understanding

A chatbot convinces several judges that it is human by using jokes and evasions. That result says something about the conversation and protocol. It does not establish that the system can verify a claim, operate safely, or perform unrelated work.

## Also called

imitation game

## Related terms

- [Artificial intelligence (AI)](https://darkfactory.dev/glossary/artificial-intelligence)
- [Artificial general intelligence (AGI)](https://darkfactory.dev/glossary/artificial-general-intelligence)

## Evidence and further reading

- [Computing Machinery and Intelligence](https://academic.oup.com/mind/article/LIX/236/433/986238)