ONLINEAGENT_OPS 2026.Q3 HOME HUBSMODELSBUILDINGCOMPANIESFILMMAKINGSECURITYLIVING WITH ITSKILLSALL HUBS, ONE TERMINAL → ARTICLES CRAFT RECORD BLOG TOOLS SKILLS SEARCH /
HOME›AI MEANINGS›WHAT IS THE TURING TEST?
AI MEANINGS · MEANINGBEGINNER

What is the Turing test? Can a machine pass as human?

The Turing test asks whether a machine can chat so like a person that a judge cannot tell. What Turing’s 1950 game shows, and what it does not.

READ5 min
WORDS1,057
SECTIONS7
TYPEMEANING
CHECKED5 OCT 26

The Turing test means a test of whether a machine can chat so much like a person that a judge cannot reliably tell which is which. The British mathematician Alan Turing proposed it in 1950, calling it the “imitation game”. Passing it shows a machine can sound human; it does not show the machine understands or thinks.

◈ IN PLAIN TERMS

Imagine texting two strangers at once. One is a person, one is a computer, and you have five minutes of questions to work out which is which. If the computer fools you about as often as it gets caught, it has done well in the Turing test. Notice what you were judging: how human the replies felt, not whether they were true.

In a sentence: “Our office ran a playful Turing test: we chatted with a bot and with a colleague, and tried to guess which was which.”

What the test is

The plain point: Turing swapped an impossible question, “can machines think?”, for a game that could actually be played.

The Stanford Encyclopedia of Philosophy, a respected reference written by experts, explains that Turing thought the question of whether machines can think was itself “too meaningless” to discuss. He proposed a game instead.1

The game has three players: a person, a machine and a questioner kept in a separate room. The questioner tries to tell which is the person; the machine tries to be mistaken for one.2 Turing guessed that by about the year 2000, an average questioner would have no more than a 70 percent chance of picking correctly after five minutes.3

Has AI passed the Turing test?

The plain point: in one 2025 study, yes, when the chatbot was told to play a human character.

Two researchers at the University of California San Diego ran the three-player version with five-minute chats. When told to adopt a humanlike persona, OpenAI’s GPT-4.5 was picked as the human 73% of the time, more often than the real people it was up against. ELIZA, a chat program from 1966, was picked 23% of the time. The authors called it the first evidence that any machine passes a standard three-party Turing test; the paper is a preprint, a study shared publicly before formal peer review.4

TAKEAWAY

Today’s chatbots can sound convincingly human for a few minutes. That is exactly why “it sounded like a person” does not tell you that you were talking to one, or that what it said was true.

What it does and does not show

The plain point: it measures how human the conversation seems, not what is going on inside.

  • It shows: a machine can produce text that people mistake for a person’s, under the test’s rules.
  • It does not show understanding. The best-known objection is the philosopher John Searle’s “Chinese Room” argument: following rules well enough to give the right replies need not mean understanding them. Searle argues against the claim that “appropriately programmed computers literally have cognitive states”.2 Does AI understand anything? looks at that question for today’s chatbots.
  • It does not show truthfulness. A reply can sound human and still be wrong; see hallucination meaning.
  • It does not mark general intelligence. Some thinkers want harder tests, such as one that also covers seeing and moving in the world.2 AGI meaning explains why no single test settles that.

Where the word comes from

The test is named after the British mathematician Alan Turing, who described it in his 1950 paper “Computing Machinery and Intelligence”, in the journal Mind, where he called it the imitation game. The Stanford Encyclopedia of Philosophy says the phrase “The Turing Test” “is most properly used to refer to a proposal made by Turing (1950)”.5

Often confused with

  • A benchmark: a fixed set of test questions with right answers, scored automatically. The Turing test is a judgement by people. See benchmark meaning.
  • “Prove you are human” puzzles: the picture puzzles on websites are quick automated checks run by a computer, not Turing’s conversation game judged by a person.
  • AGI: the idea of AI as capable as people at most mental work. Passing the Turing test is not the same thing. See AGI meaning.

Questions people ask

Has ChatGPT passed the Turing test?

In the 2025 San Diego study, OpenAI’s GPT-4.5 passed when told to play a human character, while GPT-4o, used as a comparison, scored well below chance.4 It is one study, and the authors shared it as a preprint.

Who invented the Turing test?

Alan Turing, in a 1950 paper, where he called it the imitation game.5 The history of AI places it at the start of the timeline.

Learn more on this site

QUICK CHECK · 3 QUESTIONS

Pick an answer to see if you got it. Nothing is recorded.

1. What does the Turing test check?

  • Whether a machine gets hard questions right
  • Whether a judge can tell a machine’s chat from a person’s
  • Whether a machine can see and move around
Show the answer

Answer: Whether a judge can tell a machine’s chat from a person’s. The page says the test asks whether a machine can chat so much like a person that a judge cannot reliably tell which is which.

2. If a chatbot passes the Turing test, what does that show?

  • It can sound human, not that it understands
  • It truly thinks the way a person does
  • Everything it says is true
Show the answer

Answer: It can sound human, not that it understands. The page says passing shows a machine can sound human; it does not show understanding, truthfulness or general intelligence.

3. How is the Turing test different from a benchmark?

  • It uses only multiple-choice questions
  • It is run by a computer with no people
  • It is a judgement by people, not a fixed scored test
Show the answer

Answer: It is a judgement by people, not a fixed scored test. The page says a benchmark is a fixed set of questions scored automatically, while the Turing test is judged by people.

Related terms: AI bias · Synthetic data · Training data · Transformer · Vision-language model (VLM) · World model

SOURCES
  1. 1Stanford Encyclopedia of Philosophy, The Turing Test, first published 9 Apr 2003, substantive revision 4 Oct 2021, read at source 5 Oct 2026“The phrase ‘The Turing Test’ is most properly used to refer to a proposal made by Turing (1950) as a way of dealing with the question whether machines can think. According to Turing, the question whether machines can think is itself ‘too meaningless’ to deserve discussion”. ↑
  2. 2Stanford Encyclopedia of Philosophy, The Turing Test, read at source 5 Oct 2026“Suppose that we have a person, a machine, and an interrogator. The interrogator is in a room separated from the other person and the machine. The object of the game is for the interrogator to determine which of the other two is the person, and which is the machine.” ↑a“John Searle argues against the claim that ‘appropriately programmed computers literally have cognitive states’ (64). Clearly enough, Searle is here disagreeing with Turing’s claim that an appropriately programmed computer could think.” ↑b“There are authors who have suggested that The Turing Test should be replaced with a more demanding test of one kind or another.” … on one such proposal: “the appropriate goal for research in AI has to be to construct a robot with something like human sensorimotor capabilities.” ↑c
  3. 3Stanford Encyclopedia of Philosophy, The Turing Test, read at source 5 Oct 2026, quoting Turing (1950)“I believe that in about fifty years’ time it will be possible to programme computers … to make them play the imitation game so well that an average interrogator will not have more than 70 percent chance of making the right identification after five minutes of questioning.” ↑
  4. 4Jones and Bergen (UC San Diego), Large Language Models Pass the Turing Test, arXiv:2503.23674, submitted 31 Mar 2025, read at source 5 Oct 2026; the full text lists the models as GPT-4.5 (OpenAI), LLaMa-3.1-405B, GPT-4o and ELIZA (Weizenbaum, 1966): “When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant.” … “baseline models (ELIZA and GPT-4o) achieved win rates significantly below chance (23% and 21% respectively). The results constitute the first empirical evidence that any artificial system passes a standard three-party Turing test.”↑a↑b
  5. 5Stanford Encyclopedia of Philosophy, The Turing Test, first published 9 Apr 2003, substantive revision 4 Oct 2021, read at source 12 Oct 2026bibliography: “Turing, A., 1950, ‘Computing Machinery and Intelligence,’ Mind, 59 (236): 433–60”; the game is “The Imitation Game”; and “The phrase ‘The Turing Test’ is most properly used to refer to a proposal made by Turing (1950).” ↑abibliography: “Turing, A., 1950, ‘Computing Machinery and Intelligence,’ Mind, 59 (236): 433–60”; the game is “The Imitation Game”; and “The phrase ‘The Turing Test’ is most properly used to refer to a proposal made by Turing (1950).” ↑b

Jones and Bergen, Large Language Models Pass the Turing Test, arXiv:2503.23674, 31 Mar 2025, read at source 5 Oct 2026.

Not read at source: A. M. Turing, “Computing Machinery and Intelligence”, Mind, October 1950. Turing’s words on this page are quoted from the Stanford Encyclopedia entry above.

The comparison on this page (texting two strangers) is this site’s own way of explaining, not a claim from these sources.

Free printable AI cheat sheets, and AI news every week.

Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.

SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK
aipromptgeneer://$ about this siteAI PROMPT GENEERAI explained in plain English. Every claim links to its source, with the date it was checked.NEWSLETTERWhat changed in AI each week, in plain English.SUBSCRIBE FREE ↗AI PROMPT GENEER · MACHINE ASSISTANCE DISCLOSEDEVERY ARTICLE CARRIES ITS OWN CHECKED DATE