Turing test
A test of a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human.
Historical perspective: Developed by Alan Turing (1912–1954) in 1950, he was an English mathematician, logician, and cryptanalyst who is widely regarded as the father of theoretical computer science and modern artificial intelligence.
AI agents and chatbots have passed the Turing test in multiple milestones, most notably in June, 2014, and in a landmark 2025–2026 scientific study evaluated with large language models.
On June 7, 2014 (Eugene Goostman), a chatbot impersonating a 13-year-old Ukrainian boy convinced 33% of human judges during 5-minute keyboard conversations at the University of Reading. Critics and AI researchers disputed whether this counted as a genuine pass due to the persona constraints and short duration.
In mid-2025 (GPT-4.5 and Llama-3.1-405B), in a controlled scientific study published by researchers from UC San Diego, modern large language models were tested against human participants. GPT-4.5 (when prompted with a humanlike persona) was judged to be human 73% of the time, surpassing the baseline recognition rate of actual human participants. Llama-3.1-405B was judged to be human 56% of the time, clearing the statistical threshold of indistinguishability.
"Passing the Turing test" does not mean an AI is sentient or possesses general intelligence. Instead, it measures how effectively a machine can mimic human conversational traits, tone, and fallibility in text exchanges.
Edit Word