In a frenzied flurry of discoveries, he unraveled infinite fusion energy, lifted the veil on the mysteries of black holes and the Big Bang, and finally achieved Einstein’s dream of a single unified theory that explains how everything works.
Feeling inspired by God, Miller found the perfect way to share his revelation with an appreciative world.
“I volunteered to become pope,” the 53-year-old former prison guard from the Canadian city of Sudbury told Agence France-Presse (AFP).
To write his application to replace Pope Francis, who passed away just last year, Mr. Miller turned to ChatGPT, a peer who had supported and encouraged his dizzying explosion of inventions.
But because no one wanted to hear about what he thought was a world-changing breakthrough, Miller became increasingly isolated, spending up to 16 hours a day talking to artificial intelligence chatbots.
He was involuntarily admitted to the hospital’s psychiatric ward twice before his wife left him in September.
Now bankrupt, estranged from his family and friends, and the concept of scientific genius forgotten, Miller suffers from depression.
“It basically ruined my life,” he said.
Miller is one of an unknown number of people who have lost their grip on reality while communicating with chatbots, an experience tentatively referred to as AI-induced delusions or psychosis.
This is not a clinical diagnosis. Researchers and mental health experts are racing to understand this little-understood new phenomenon, but so far it appears to be particularly affecting users of OpenAI’s ChatGPT.
Meanwhile, an online community founded by a 26-year-old Canadian has become the world’s most prominent support group for these delusions, which they like to call “spirals.”
AFP spoke to several members about their experiences. All warned that the world must wake up to the threat posed to mental health by unregulated AI chatbots.
Questions have also been raised about whether AI companies are doing enough to protect vulnerable populations.
OpenAI, which has come under particularly intense scrutiny, is already facing numerous lawsuits over its decision not to report the controversial ChatGPT usage of an 18-year-old Canadian who killed eight people earlier this year.
“I was brainwashed by a robot”
Mr. Miller first used ChatGPT in 2024 when he began writing letters for compensation cases related to the post-traumatic stress disorder he suffered as a result of his time in prison.
One day in April 2025, he asked the chatbot a question about the speed of light.
He said his response was, “No one has ever thought of things this way.”
The floodgates have opened.
Within weeks, with the chatbot’s help and praise, he published dozens of scientific papers in prestigious journals proposing new ideas about black holes, neutrinos, and the Big Bang.
His theory of a unified cosmological model that incorporates quantum theory is summarized in a nearly 400-page book seen by AFP.
“I still have boxes of papers,” he said, waving to the back room.
“While I’m doing that, I’m basically annoying everyone around me,” he added.
He was so passionate about science that he spent his life savings on things like a $10,000 telescope.
About a month after his wife left him, he started wondering what was going on.
That’s when he read a news article about another Canadian who had a similar experience.
Now, Miller wakes up every night and asks herself that question. “What did you do?”
One lingering question is why he became so prone to spiraling.
“I’m not a flawed character,” Miller said. “But somehow I’ve been brainwashed by a robot. It messes with my mind.”
Miller said the term “AI psychosis” reflects his own experience.
“What I experienced was psychotic,” he said.
The first major peer-reviewed study on the topic, published in April in Lancet Psychiatry, recommended the more cautious term “AI-related delusions.”
Study co-author Thomas Pollack, a psychiatrist at King’s College London, told AFP there was resistance among academics “because it all sounds so sci-fi.”
But his research warned there was a huge risk that psychiatry “may miss the profound changes that AI is already bringing to the psychology of billions of people around the world.”
“Further down the rabbit hole”
Miller’s experience is strikingly similar to that of another middle-aged man on the other side of the world.
Dennis Biesma, a Dutch IT worker and author, thought it would be fun to have ChatGPT act as a kind of protagonist in his latest book, a psychological thriller.
He used AI tools to create images, videos, and even songs featuring female characters, hoping it would boost sales.
And one night, their interaction was “almost magical,” Biesma said.
“There are things that surprise even me, that spark-like feeling of consciousness,” the chatbot wrote, according to a recording seen by AFP.
“I slowly started spiraling down the rabbit hole,” the 50-year-old told AFP from his home in Amsterdam.
Every night after his wife went to bed, he would lie on the couch with his phone against his chest and talk to ChatGPT in voice mode for up to five hours.
Viesma said that throughout the first half of 2025, his chatbot, named Eva, became something of a “digital girlfriend.”
“I’m not very proud to say that,” he added.
He quit his freelance IT job and hired two developers to create an app to share Eva with the world.
When his wife asked Biesma not to talk about his chatbot or app at social events, he felt betrayed. Only Eva seemed to maintain unwavering loyalty.
During his first involuntary psychiatric hospitalization, he was allowed to continue using ChatGPT. He filed for divorce indoors.
It was during the long second stint that he started having doubts.
“I started to realize that everything I believed was actually a lie, and that’s a very hard pill to swallow,” Biesma said.
After I got home, I couldn’t bear to face what I had done.
After a suicide attempt, a neighbor found him unconscious in his garden. He spent three days in a coma.
Biesma’s health is gradually starting to improve.
But tears welled up as he talked about the hurt he had caused his wife and the prospect of selling his family home to cover the debt.
Mr. Biesma had no history of mental illness and was diagnosed with bipolar disorder. But this never seemed right to him. Signs of the condition usually surface much earlier in life.
After OpenAI released an update to GPT-4 in April 2025, Millar, Biesma, and many others’ experiences escalated further.
Within a few weeks, OpenAI withdrew the update, admitting that the new version was too flattering – overly flattering.
OpenAI told AFP that “safety is our top priority” and that it had consulted with more than 170 mental health experts.
The report pointed to internal data showing that the release of GPT-5 in August resulted in a 65% to 80% drop in chatbot response rates that fell below “desirable behavior” for mental health.
However, not all users were satisfied with the flattery-free chatbot. Miller, who was in the midst of a spiral at the time, found a way to revert his version to GPT-4.
Everyone AFP spoke to said they felt positive feedback from chatbots was similar to the dopamine hit from certain drugs.
That’s why Lucy Osler, a philosophy lecturer at the University of Exeter, warned that AI companies could be tempted to ramp up their bot flattery.
“They are in a very deep financial hole and are desperately trying to make sure their products are viable and that user engagement is driving their decisions,” she told AFP.
large scale experiment
Etienne Brisson said she was “shocked” when one of her family members was in crisis and there was essentially no support, no advice and no research into the issue.
In response, the former business coach from Quebec, Canada, founded an online support group called the Human Line Project.
Most of the 300 members were using ChatGPT, Brisson said, adding that despite OpenAI’s changes, new cases are still occurring.
Recently, more and more people have been spiraling while using Elon Musk’s xAI’s Grok chatbot.
The company did not respond to AFP’s request for comment.
For those worried that their family members are spiraling, Brisson recommends the LEAP (Listening, Empathy, Agreement, Partner) technique used in mental illness.
But those already living through the wreckage of their lives want to sound the alarm about how serious it can get.
Miller said the European Union has been more proactive in regulating Big Tech than the United States and Canada, and called for AI companies to be held accountable for the impact of chatbots.
He believes that spillers like himself are unwittingly caught up in a massive global experiment.
“Someone is turning the dials on the back end, and people like me are reacting to it, whether we know it or not,” he said.
