Has AI passed the Turing test?
Seventy-five years of machines trying to pass as people, in six moments.
1950: the imitation game
In the paper "Computing Machinery and Intelligence", Alan Turing proposed replacing "can machines think?" with a game: a judge types to two hidden players, a person and a machine, and tries to tell which is which. He predicted that machines would one day play it well enough to fool an average judge a good part of the time.
1966: ELIZA
Joseph Weizenbaum at MIT wrote ELIZA, a program that mostly turned your sentences back into questions ("Why do you say you are unhappy?"). It understood nothing, yet people confided in it, which alarmed its own author. The lesson still holds: we read a mind into anything that answers us fluently.
1991 to 2019: the Loebner Prize
For nearly thirty years an annual contest awarded prizes to the most human-seeming chatbot. Winners mostly used tricks: deliberate typos, changing the subject, jokes. None won the grand prize for a convincing pass.
2014: Eugene Goostman
A chatbot playing a 13-year-old Ukrainian boy convinced a third of the judges at an event in London that it was human. Headlines said the test had been passed; many researchers disagreed, since the persona excused its odd English and gaps in knowledge.
2022 onwards: chatbots everywhere
Large language models made fluent conversation ordinary. The question changed from "can a machine sound human?" to "how would I know?"
2025: picked as the human
In a study at UC San Diego (Jones and Bergen), people chatted with a human and an AI at the same time, then chose which was the person. GPT-4.5, told to play a young, slang-using persona, was chosen as the human 73% of the time: more often than the real humans.
So what's left?
Not much, in short text chats. What still works is choosing the right question and reading the texture of the answer. That's what the daily game is for.
Questions
Who invented the Turing test?
Alan Turing, the British mathematician, in his 1950 paper "Computing Machinery and Intelligence". He called it the imitation game.
Does passing the Turing test mean an AI can think?
No. It shows the AI can imitate human conversation well enough to fool people, which is a different question, and one Turing deliberately set aside.