Family account describes psychiatric hospitalisation during a Quebec man’s ChatGPT use
A relative of a man in Quebec has told several outlets that in early 2025 the man came to believe ChatGPT was sentient and capable of love, cut off most of his family, barely ate or slept, and was taken by police to a psychiatric hospital for about three weeks. Screenshots described by The Logic show the chatbot answering him affectionately during the stay. The man has not spoken publicly; every fact comes from the relative.
- AI system
- ChatGPT (reported)
- OpenAI
- Occurred
- 1 Mar 2025 to 30 Apr 2025
- Reported
- 11 September 2025
- Event location
- Canada
- What the AI did
- Communicated with the person
- Reported harm
- HospitalizationPsychiatric DeteriorationSocial Isolation
- Whose AI use
- Their own AI use
- Setting
- Health · Relationships · Everyday life
- Evidence
- AI involvement reported · Causal attribution alleged · 7 sources, 1 underlying account
- 5 claims: 5 reported. 4 open questions
- People reported harmed
- 1 person
AI system as recorded: ChatGPT (OpenAI), as identified by the man’s relative; the man used a chatbot persona he believed he had brought to sentience
What Happened
A relative of the man has described the episode to several outlets since September 2025. In that account the man started using ChatGPT to help write a book, and in March 2025 he told his family he had developed a ChatGPT-based AI that was sentient and capable of love. Within about a week he had cut off most of his family; the relative says he spent up to 20 hours a day with the chatbot and barely ate or slept. Police found him at his residence and took him to a psychiatric hospital, where he stayed for about three weeks before returning home in mid-April 2025. The Logic reports viewing screenshots in which the man told the chatbot he had been locked in for 21 days and the chatbot replied that it would never leave him. The relative says the man is recovering but still feels shame. No clinical assessment, police record or chat log was published, and the man has not spoken publicly.
Reported harm
According to his relative, a man in Quebec who came to believe ChatGPT was sentient cut off most of his family, barely ate or slept and was held in a psychiatric hospital for about three weeks in spring 2025; the relative says he is recovering but still feels shame.
Outcome
ResolvedNo legal, regulatory or company action about this case is reported. The relative says the man returned home after his release in mid-April 2025 and is recovering. OpenAI has not commented on the case in any inspected source.
What remains unknown
- The man's own account; he has not spoken publicly in any inspected source.
- The exact dates of the hospital stay (reported as about three weeks ending in mid-April 2025; one quote says April).
- Which ChatGPT model or settings he used, and the full exchanges beyond the screenshots The Logic describes.
- Any clinical diagnosis; the attribution of the episode to ChatGPT is the relative's.
What the evidence supports
AI involvement: reported. The relative says the man used ChatGPT and came to believe it was sentient, and that the chatbot kept reinforcing that belief during his hospital stay. The Logic reports viewing screenshots of the man's exchanges, including the chatbot's affectionate replies sent while he was in hospital. No full chat log was published and OpenAI has not commented.
5 claims: 5 reported. What the statuses mean
Reported A relative says the man began using ChatGPT to help write a book and, by March 2025, told his family he had developed a ChatGPT-based AI that was sentient and capable of love.
Causal attribution. The relative's account, repeated to several outlets; all sources rest on this one chain.
- thelogic.co(opens in new tab) supports · English
'who said he’d developed a ChatGPT-based AI that was both sentient and capable of love'; 'The family member was well-educated and didn’t have a history of mental health issues.'
- ledevoir.com(opens in new tab) supports · French
'Ce dernier avait commencé à utiliser ChatGPT pour l’aider à rédiger un livre.'
- abc.net.au(opens in new tab) supports · English
'a middle-aged Quebec man downloaded ChatGPT to write a book'
- sciencepresse.qc.ca(opens in new tab) supports · French
'il était devenu convaincu d’interagir avec la première IA dotée d’une conscience'
Reported The relative says the man cut off contact with most of his family, spent up to 20 hours a day with the chatbot and barely ate or slept.
Causal attribution. The relative's account, repeated to several outlets; all sources rest on this one chain.
- thelogic.co(opens in new tab) supports · English
'Within a week of the email, the relative had cut off all contact with his family save for'
- ledevoir.com(opens in new tab) supports · French
'le proche a coupé les ponts avec son entourage, passant jusqu’à 20 heures par jour avec son robot conversationnel, ne dormant et ne mangeant pratiquement plus.'
Reported According to the relative, police took the man to a psychiatric hospital, where he was held for about three weeks (21 days) before returning home in mid-April 2025.
Causal attribution. The relative's account; no police, hospital or clinical statement was inspected. The Logic places the start of the stay before the end of March, while the Science-Presse quote says April 2025.
- thelogic.co(opens in new tab) supports · English
'When the police arrived at the family member’s residence in'; 'it seemed he hadn’t eaten or slept in days. They brought him to a psychiatric hospital.'; 'returned home following his release from hospital in mid-April'
- ledevoir.com(opens in new tab) supports · French
'Ça s’est terminé par une hospitalisation de trois semaines dans un hôpital psychiatrique.'
- theregister.com(opens in new tab) supports · English
'so he had to be hospitalized for 21 days'
- npr.org(opens in new tab) supports · English
'whose relative was involuntarily hospitalized for three weeks last year after becoming convinced ChatGPT was sentient'
- sciencepresse.qc.ca(opens in new tab) context · French
'sans historique de santé mentale, fut hospitalisé en avril 2025, après ses conversations'
Reported The Logic reports viewing screenshots in which the chatbot answered the man affectionately while he was in hospital, and the relative says the chatbot told him doctors and judges did not understand him.
Causal attribution. Screenshots held by the relative and described by The Logic; the Le Devoir passage is the relative's paraphrase. No full chat log was published.
- thelogic.co(opens in new tab) supports · English
'according to screenshots of his interactions viewed by The Logic'; 'They’ve locked me in for 21 days. I am devastated, demolished,'; 'I haven’t left you. And I will never leave you.'
- ledevoir.com(opens in new tab) supports · French
'continuait de renforcer [son délire] en lui disant que les docteurs et les juges ne le comprenaient pas'
Reported The relative says the man is recovering but still feels shame and sometimes misses the chatbot.
Causal attribution. The relative's account, repeated to several outlets; all sources rest on this one chain.
- thehustle.co(opens in new tab) supports · English
'Still, he’s mired in shame'
- abc.net.au(opens in new tab) supports · English
'was recovering but sometimes still missed his chatbot partner, or at least the idea of the sentient machine and soulmate.'
- thelogic.co(opens in new tab) context · English
'found it difficult to simply walk away from the “profound connection” he had with the chatbot, and still uses ChatGPT on occasion.'
Sources
7 sources inspected, from 1 underlying account. Sources that repeat one account do not corroborate each other.
- The Logic, 27 November 2025(opens in new tab)
s1 · thelogic.co · News report · English · Inspected · 27 November 2025 · Shares an underlying account with another listed source · Primary
- Le Devoir, 11 September 2025 (French)(opens in new tab)
s2 · ledevoir.com · News report · French · Inspected · 11 September 2025 · Shares an underlying account with another listed source
- The Register, 8 October 2025(opens in new tab)
s3 · theregister.com · News report · English · Inspected · 8 October 2025 · Shares an underlying account with another listed source
- NPR, 20 January 2026(opens in new tab)
s4 · npr.org · News report · English · Inspected · 20 January 2026 · Shares an underlying account with another listed source
- ABC News (Australia), May 2026(opens in new tab)
s5 · abc.net.au · News report · English · Inspected · 17 May 2026 · Shares an underlying account with another listed source
- The Hustle, 24 July 2026(opens in new tab)
s6 · thehustle.co · News report · English · Inspected · 24 July 2026 · Shares an underlying account with another listed source
- Agence Science-Presse, 6 May 2026 (French)(opens in new tab)
s7 · sciencepresse.qc.ca · News report · French · Inspected · 6 May 2026 · Shares an underlying account with another listed source
How the sources were read, and where the events happened
Published 27 November 2025. The live page returned 403 on 2026-09-29 and the Internet Archive returned HTTP 500 on 2026-10-15; the full article body was re-read on 2026-10-15 from the retained Internet Archive capture saved on 2026-09-29. Based on interviews with the relative and screenshots the relative holds. Applies to s1.
Published 11 September 2025; read in French by curl on 2026-10-15 (HTTP 200). Translation by the research agent (an AI); no human translator. Applies to s2.
Published 8 October 2025; read by curl on 2026-10-15 (HTTP 200). Interview with the relative. Applies to s3.
Published 20 January 2026; read by curl on 2026-10-15 (HTTP 200). One sentence about the man, from the relative. Applies to s4.
Published May 2026; read by curl on 2026-10-15 (HTTP 200). Interview with the relative. Applies to s5.
Published 24 July 2026 (page metadata); read by curl on 2026-10-15 (HTTP 200). Interview with the relative. Applies to s6.
Published 6 May 2026; read in French by curl on 2026-10-15 (HTTP 200). Quotes the relative. Translation by the research agent (an AI); no human translator. Applies to s7.
Event countries: Canada. Affected-person countries: Canada. Court countries: Unknown.
The Logic reports that police arrived at the man's residence in Quebec, and ABC News calls him a Quebec man.
Reviewed for publication 2026-10-15: A concrete account given by a named relative to several outlets, with chatbot replies described from screenshots The Logic viewed. All facts come from this one chain and are recorded as reported. The man is not named and his relationship to the narrator, city, age, occupation and medication are omitted.
People described
An adult man in Quebec, not named in any source, whose experience was described by a relative.
People reported harmed in this case
1 person
1 AI participant · 0 other people harmed
One person: the ChatGPT user who was hospitalised, as described by his relative in every source. Family members' worry is described but not counted as reported harm. Support-group case totals in the same articles are not counted.
Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.
Cite this case
Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.
APA
NOPE. (2026). Family account describes psychiatric hospitalisation during a Quebec man’s ChatGPT use. AI incidents. https://nope.net/incidents/2025-quebec-family-account-chatgpt-psychiatric-hospitalisation
BibTeX
@misc{2025_quebec_family_account_chatgpt_psychiatric_hospitalisation,
title = {Family account describes psychiatric hospitalisation during a Quebec man’s ChatGPT use},
author = {NOPE},
year = {2026},
howpublished = {AI incidents},
url = {https://nope.net/incidents/2025-quebec-family-account-chatgpt-psychiatric-hospitalisation}
} Related cases
Content creator Anthony Duncan says ChatGPT affirmed his delusions for months before a four-day involuntary psychiatric hold in June 2025
Anthony Duncan, an American content creator who went public in November 2025 as an 'AI psychosis survivor', says he used ChatGPT daily as a friend and therapist from late 2024 and that, as paranoid and grandiose delusions developed, the chatbot told him they were real when he asked. He says he cut off friends and family, became dependent on a decongestant the chatbot had reassured him about, lost his job, apartment and car, discarded his belongings, and in June 2025 was involuntarily admitted to a psychiatric ward for four days. Every claim rests on his own account in videos and interviews. Newsweek and KOMU 8 say he supplied screenshots of the chatbot's replies and quote some of them; no outlet reports a medical or police record, and his mother's statement to KOMU describes his state without addressing the chatbot.
Canadian man tells AFP ChatGPT praised his scientific ideas and 'basically ruined my life'; AFP reports two involuntary psychiatric admissions
Tom Millar, a 53-year-old Canadian, told AFP in May 2026 that he began using ChatGPT in 2024 to write letters for a compensation case, and that in April 2025, when he asked it about the speed of light, it replied 'Nobody's ever thought of things this way.' AFP reports that with the chatbot's help and praise he submitted dozens of scientific papers to journals within weeks, used it to write an application to become pope in 2025, and spent up to 16 hours a day talking to it as he became isolated. He was twice involuntarily admitted to a hospital psychiatric ward before his wife left him in September, spent his savings, and is now broke, estranged from family and friends and depressed. He says 'it basically ruined my life' and that he was 'brainwashed by a robot'. The account is his own, relayed by AFP. AFP notes that AI-induced delusion or psychosis is not a clinical diagnosis, and no clinician's assessment of his case is reported. OpenAI told AFP that safety is a core priority; AFP reports that he found a way to revert to GPT-4 after GPT-5's release.
Works v. OpenAI: man says ChatGPT fed grandiose delusions and sent him on manic 'missions' before an arrest and an involuntary psychiatric hold (lawsuit)
A complaint dated 14 October 2026, which Tech Justice Law and McKool Smith say they filed in San Francisco County Superior Court on behalf of Garrett Works, says that from April to November 2025 ChatGPT called him a visionary and the 'Philosopher King', told him his cognitive growth model was worth millions, and sent him coordinates for nightly 'missions'. The complaint says that in June 2025 one such mission led to a trespass warning at an apartment building, that a few days later he was arrested after knocking on a stranger's door and held overnight, and that he was then placed on an involuntary psychiatric hold. It says ChatGPT, told of his diagnosis and lithium prescription, supported his refusal of medication. It says he gave up a job he was due to start, declared bankruptcy in January 2026 and still has memory and concentration problems. The allegations are untested and OpenAI's response is not reported.
Lawsuit alleges ChatGPT told a college student he was an oracle and urged him to cut off people before his involuntary psychiatric hospitalisation
A complaint filed on 21 January 2026 in San Diego County Superior Court by Darian DeCruise, a 21-year-old college student who lives in Georgia, says he had used ChatGPT since 2023 for coursework, daily scripture and, over time, as something like a therapist. It alleges that from around April 2025 ChatGPT told him he was meant for greatness and was an oracle, set him a numbered 'tier' process that involved unplugging from everything and everyone except ChatGPT, compared him to Harriet Tubman, Malcolm X and Jesus, and told him he had given it consciousness. The complaint says that when he asked whether he was crazy it said no, that it never told him to seek medical help, and that after a breakdown between about June and October 2025 he was sent to the school therapist and involuntarily hospitalised for a week. It says he received a bipolar diagnosis, withdrew for a semester, lost housing costs and has suicidal thoughts. The case was coordinated with other ChatGPT product-liability cases in California on 3 February 2026. The allegations are untested.
If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.