The Turing test: can a machine pass for a human?
The Turing test is an experiment proposed by British mathematician Alan Turing in 1950: a judge chats in writing with a human and a machine and has to work out which is which. If the machine fools the judge, it “passes”, but that does not prove it can think.
Data as of October 22, 2026
Key figures
| Level | Beginner |
|---|---|
| Origin | 1950 |
| Introduced by | Alan Turing, « Computing Machinery and Intelligence », revue Mind |
How it works
Picture three people in separate rooms. A judge types questions. Two players answer in writing: one is a human, the other a machine. At the end, the judge has to say which one is the machine. If the machine fools the judge about as often as a human would, it passes.
Alan Turing set out this idea in 1950 in “Computing Machinery and Intelligence”, published in the journal Mind (original paper). Instead of arguing endlessly over whether machines can think, he proposed a practical experiment he called the “imitation game”. It is often cited as one of the founding moments of artificial intelligence, and you can follow what came next in the AI timeline.
Why it is back in the news
For decades, chat programs gave themselves away quickly. With large language models, chatbots now write in a strikingly natural way. In March 2025, Cameron Jones and Benjamin Bergen published a study in which, during five-minute conversations, judges picked GPT-4.5 as the human more often than the actual human participants (research paper on arXiv).
In everyday terms, this means a message you receive by email, in a customer support chat or on social media may have been written by AI, and you probably cannot tell from the style alone.
What the test does not measure
This is where the most common misconceptions come in:
- “Passing the test means thinking like a human.” The test measures how things look in a short conversation. An AI can be convincing and still make reasoning mistakes or produce hallucinations.
- “The Turing test is the official benchmark for general AI.” There is no official benchmark. The idea of artificial general intelligence covers far broader abilities than conversation.
- “People can always spot a machine.” Recent research shows that is not true under some conditions, so gut feeling is not enough.
The takeaway
How smoothly a conversation flows tells you nothing about whether the content is accurate or who is on the other end. If an online exchange pushes you to act fast, whether that means paying, clicking or sharing personal details, verify through another channel, whether the sender is a person or a machine.
Related terms
- Artificial intelligence · Techniques that let machines perform tasks we associate with human intelligence, such as understanding, predicting or creating.
- Chatbot · A program you talk to in writing or by voice, which replies in everyday language like a conversation partner.
- Artificial general intelligence (AGI) · A hypothetical AI able to match humans across most intellectual tasks; there is no single agreed definition or timeline.
- Benchmark · A standardised test, a fixed set of scored questions or tasks, used to compare how different AI models perform.
- Large language model (LLM) · A program trained on huge amounts of text to predict what comes next in a text; it is the engine behind AI assistants.
Try it
Full glossary
Frequently asked questions
Has ChatGPT passed the Turing test?
A study published in March 2025 by researchers Cameron Jones and Benjamin Bergen pitted participants against a human and an AI in five-minute conversations. GPT-4.5, prompted to adopt a humanlike persona, was picked as the human more often than the real humans were. It is a result in a specific setup, not proof of human-level intelligence.
If an AI passes the Turing test, is it conscious?
No. The test measures how convincingly a system imitates human conversation, not consciousness, understanding or feelings. Turing himself proposed the game to replace the question “Can machines think?”, which he considered too vague.
Why is it also called the imitation game?
That is the name Turing used in his 1950 paper. The 2014 film about his life borrowed the same title.
Do AI researchers still use the Turing test?
It remains a cultural reference point and still inspires experiments, but models are mostly evaluated with benchmarks: sets of specific tasks in math, coding or reasoning whose scores can be compared.
Data sources
Data as of October 22, 2026
Report an error
Does a figure, a date or a fact look wrong? Tell us: every report is checked.