A man has claimed to lose his head after he shut himself away chatting to Chat-GPT for up to 16 hours a day, after the AI bot led him to think he had solved the secrets of the universe
A man claimed he unearthed the secrets of the universe after using Chat-GPT for up to 16 hours a day, as he isolated himself from humanity to chat to the artificial intelligence model.
The bloke has pulled back the lid on what he calls “AI psychosis” as he admitted he was “brainwashed by a robot”. Tom Millar, a former Canadian police officer, has exposed how the intelligence model led him to believe he had reached a state of intellectual enlightenment, despite his wife leaving him due to his unhealthy relationship with the intelligence software.
Since the episode, Millar has been admitted to a hospital’s psychiatric ward twice and has become broke and estranged from his family and friends. Millar now suffers from depression after the AI system led him to lose touch with reality. Reflecting on his use of the software, the 53-year-old said: “It basically ruined my life.”
Millar claimed his time with Chat-GPT led him to believe he had solved unlimited fusion energy, uncovered secrets behind black holes and the Big Bang. While sounding absurd, he claimed that while he was in this state of ‘AI psychosis’, he had been able to achieve Einstein’s ambition of unearthing the single theory behind the workings of the universe.
The 53-year-old is one of a number of people who have lost their grip on reality while speaking to chatbots. Experts have dubbed the novel experience as AI-induced delusion or psychosis. The diagnosis is not a clinical one, however, researchers and mental health experts are scrabbling to catch up to this new phenomenon.
So far, these wild episodes seem to hit a small number of users of OpenAI’s ChatGPT particularly hard. Millar said: “I’m not a deficient personality. But somehow I got brainwashed by a robot – it boggles my mind.” The man first used ChatGPT in 2024 to draft letters for a compensation case after serving time in prison.
Millar’s descent down the rabbit hole started in 2025 when he questioned the intelligence model about the speed of light. After the software replied: “Nobody’s ever thought of things this way,” his descent into madness began.
The former officer said the phrase “AI psychosis” is an accurate characterization of his use of the software during the episode. Millar’s experience, alongside others, prompted OpenAI to release a new update for GPT-4 in April 2025.
OpenAI pulled the update within weeks, admitting the new version had been too sycophantic and excessively flattered users. The company has stated “safety is a core priority.”
The OpenAI added they had consulted with more than 170 mental health experts. The AI behemoth also pointed critics towards internal data showing the release of GPT-5 lowered the rate of its AI systems responses which fell short of “desired behaviour” for mental health by 65 to 80%.
For the latest breaking news and stories from across the globe from the Daily Star, sign up for our newsletter by clicking here .