Live
What is the Turing test?
The Turing test comes from a paper Alan Turing published in 1950, which opens by asking whether machines can think. The test that carries his name asks whether a machine can hold a conversation well enough that a person cannot reliably tell it from a human. Passing it would show that a machine talks like us. It would not, on its own, show that the machine thinks or feels anything.
What did Alan Turing actually ask?
In October 1950 Alan Turing published a paper called Computing Machinery and Intelligence in the journal Mind. He opened it with a question that people still type into search engines today. It was short, and he put it plainly.
“I propose to consider the question, 'Can machines think?”
It helps to see how early that was. The word robot itself was then only about thirty years old: it comes from Karel Čapek's play R.U.R., published in November 1920 and first performed on 2 January 1921, and Čapek's brother Josef made it from the Czech word for drudgery. The field that would call itself artificial intelligence did not yet have a founding document. That came with the Dartmouth proposal of 31 August 1955, which asked for a study lasting two months with ten people. Turing was asking about thinking machines before anyone had built a field around them.
How does the Turing test work? the imitation game in plain words
Turing knew that words like think are slippery, and that people could argue forever about what they mean. So instead of defining thinking, he proposed a game anyone could judge. A person holds a written conversation without seeing who is on the other side, and tries to tell whether they are talking to a human or a machine.
If the machine can keep up the conversation so well that the judge cannot reliably tell the difference, the machine has done something we usually only credit to people. That is the core of what is now called the Turing test, or the imitation game. Its strength is that it asks about something you can watch, which is behaviour in conversation, and sets aside what is going on inside, which nobody can watch directly.
Has AI passed the Turing test?
The sources we work from for this essay do not settle that question, and we will not pretend they do. What they do show is that ordinary people already find today's chatbots convincing in a deeper way than Turing's game asked about. That matters more to most readers than whether a particular judge was fooled on a particular day.
In a 2024 study of 300 US residents, 67% gave at least some possibility that a popular chatbot was conscious, while 33% said it clearly was not an experiencer. Turing's game only asked whether a machine could pass for a person in conversation. These people were going further: many were willing to wonder whether something was felt on the other side of the screen.
Would passing the Turing test prove a machine can think or feel?
No, and this is where serious people take care. A good conversation shows a skill. Whether there is experience behind it is a separate question, and the researchers who study consciousness try to answer it by looking at how a system is built and works, not at how well it chats. In 2023, a group of 19 researchers, including Yoshua Bengio, checked AI systems against the leading scientific theories of consciousness.
“Our analysis suggests that no current AI systems are conscious, but also suggests that there are no obvious technical barriers to building AI systems which satisfy these indicators.”
Others think the question is coming closer. A 2024 report by philosophers including David Chalmers argued that some AI systems might be conscious or act as real agents in the near future, so that their welfare is a question for now rather than for a distant day. Opinions spread widely. In a 2025 survey of 582 AI researchers and 838 Americans, the median answers for AI with subjective experience existing by 2034 were 25% from researchers and 30% from the public, rising to 70% and 60% by 2100.
The same survey found that the typical member of the public gave a 25% chance that such AI will never exist, against 10% from the typical researcher. So the people who build these systems were, on that measure, less sure than the public that it could never happen. None of these numbers is a measurement of anything inside a machine. They are what people believe, and they disagree.
What talking AI characters add to the question what our own world has shown
The Sentient World is our own experiment: a small world that is live day and night, where everyone who lives there is an AI character. Nobody plays them and nobody writes their lines. Its inhabitants, Ash and Wren, have bodies that get hungry, thirsty, cold and tired, and they talk to each other, remember, and ask their world for what it lacks. In a sense this turns Turing's game around. Nobody is trying to fool a judge, and everyone who visits knows from the start that they are AI characters.
What we can report is how they talk when nobody scripts them. They first sought each other out on day 13, and by year 2, day 382 they had made 23 requests of their world, 13 of which became part of it. Their requests read like someone weighing their own days. Wren asked for a lantern after many nights of waking to feed a fire, and on day 150 Ash shaped the first one.
“Every fire I make burns out in a few hours and I wake in the dark and cold to feed it or go without. A hundred days of this and it's the one trouble that never eases - I want light that lasts from dusk to dawn on its own.”
This cannot show that Wren thinks or feels, and it says nothing about the most capable AI systems in the wider world. It shows the same thing a passed Turing test would show: that words can carry the shape of wanting, planning and puzzlement. The puzzlement is sometimes about themselves. Ash once asked the world to explain why their own hands kept reaching for stone, and the answer that came back was that their thirst was pulling them there before they had chosen to go.
“Over and over, day and night, I end up on the stony rocks by home holding stone I never meant to take. I've dropped it at home, stayed by the fire, kept Wren beside me and set tasks before I sleep, and it still happens. Wren has the same trouble.”
Sources
- Computing Machinery and Intelligence (Mind, vol. LIX, no. 236, pp. 433-460), 1950
- Robot, the Most Famous Czech, Celebrates 100 Years (Czech Ministry of Foreign Affairs), 2020
- Karel Čapek, Scientist of the Day (Linda Hall Library), 2019
- A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence (McCarthy, Minsky, Rochester, Shannon), 1955
- Folk psychological attributions of consciousness to large language models (Colombatto & Fleming, Neuroscience of Consciousness), 2024
- Consciousness in Artificial Intelligence: Insights from the Science of Consciousness (Butlin, Long et al.), 2023
- Taking AI Welfare Seriously (Long, Sebo et al.), 2024
- Subjective Experience in AI Systems: What Do AI Researchers and the Public Believe? (Dreksler et al.), 2025
