Skip to main content
NOPE
Critical reported severity Media Coverage

BJPsych case report describes first-episode psychosis temporally associated with months of intensive AI storytelling use, ending in serious injury

A case report by Cortney Husodo and Gurubhaskar Shivakumar in The British Journal of Psychiatry, published online on 24 April 2026, describes a woman in her early 40s, given the pseudonym Alice, with no prior psychiatric history. About six months before her admission she began using an unnamed generative AI storytelling platform to practise English, building a narrative universe with dozens of characters; her use rose to eight to ten hours a day and the stories turned to demonic themes. Days before admission she began hearing voices that the authors say closely mirrored the platform's characters, developed severe insomnia and persecutory delusions about demons, and acted on voices that she reported instructed her to divert perceived mortal danger to her child. She was found unconscious and seriously injured. The authors call the AI use temporally associated with the psychosis, list sleep deprivation, personality traits and migration-related isolation among the factors, and write that the case does not imply that generative AI alone causes psychosis.

AI system
Unidentified AI storytelling platform
Occurred
Event date unknown
Reported
24 April 2026
Event location
Unknown
What the AI did
Communicated with the person
Reported harm
Physical InjuryHospitalizationPsychiatric Deterioration
Whose AI use
Their own AI use
Setting
Health · Everyday life
Evidence
AI involvement reported · Causal attribution unclear · 2 sources, 1 underlying account
6 claims: 3 documented, 3 reported. 5 open questions
People reported harmed
1 person

AI system as recorded: Generative AI storytelling platform (not named in the case report)

Reported harm

A woman developed first-episode psychosis with voices and delusions that the authors say mirrored characters from months of intensive AI storytelling use, and was found unconscious and seriously injured after acting on those voices; the authors describe a temporal association within several other factors.

Outcome

Unknown
  • 24 April 2026: case report published online (First View) in The British Journal of Psychiatry: Husodo C., Shivakumar G., 'Alice in AI-Wonderland: digital dyad, distorted reality', DOI 10.1192/bjp.2026.10641.
  • The report describes her admission after she was found unconscious and seriously injured; it does not give her treatment course, discharge or later outcome.
  • No lawsuit, regulatory action or platform response is described. No eLetters had been published on the article when it was read on 23 October 2026, and no news coverage of the case was found.

What remains unknown

  • When the injury and admission happened; the report gives only relative timings and was published on 24 April 2026.
  • Where the patient lived and was treated.
  • Which storytelling platform she used and what its replies said; no conversation record is published and the clinical data are not public.
  • Her treatment course, recovery and the long-term effect of her injuries.
  • How far the AI use contributed to the psychosis; the authors describe a temporal association among several factors.

What the evidence supports

AI involvement: reported. The case report says that over about six months she built fictional narratives with a generative AI storytelling platform, with use rising to eight to ten hours a day and stories turning to demonic themes, and that the demonic figures and several of the voices of her psychosis corresponded directly to characters originating from those AI-generated narratives before she was found unconscious and seriously injured. The authors call the AI use temporally associated with the psychosis, list sleep deprivation, personality traits and isolation among the factors, and do not attribute the voices' instruction to the platform's output.

6 claims: 3 documented, 3 reported. What the statuses mean

Documented Cortney Husodo and Gurubhaskar Shivakumar published a case report, 'Alice in AI-Wonderland: digital dyad, distorted reality', in The British Journal of Psychiatry, online on 24 April 2026, describing a patient they call Alice.

Causal attribution. The article itself establishes its authorship, title and publication date.

Documented The case report describes a woman in her early 40s, an immigrant with no prior psychiatric history, who was found unconscious and seriously injured, and who reported acting in response to voices that instructed her so as to divert perceived mortal danger to her school-aged child; she was admitted to hospital.

Causal attribution. Clinical facts recorded by the authors; the instruction came from hallucinated voices, which the report does not attribute to platform output.

  • cambridge.org(opens in new tab) supports · English
    ‘Alice’, an immigrant woman in her early 40s with no prior psychiatric history, was found unconscious and seriously injured; She reported acting in response to voices that instructed her; so as to divert perceived mortal danger to her school-aged child.; In the days preceding hospital admission
Reported According to the case report, about six months earlier she had begun using a generative AI storytelling platform to practise English, created a character that grew into a narrative universe of dozens of interconnected characters, increased her use to eight to ten hours a day, and saw the narratives shift to fantastical content involving demonic entities, while retaining insight that they were fiction until shortly before admission.

Causal attribution. Her history as given in the case report; no platform records or conversation logs are published.

  • cambridge.org(opens in new tab) supports · English
    Approximately 6 months earlier, Alice had begun using the storytelling platform to practise English.; this evolved into a complex narrative universe populated by dozens of interconnected characters.; Over time, her engagement escalated to eight to ten hours daily.; The narratives gradually shifted from realistic interpersonal themes to increasingly fantastical content involving demonic entities.; Until shortly before her admission, she retained insight that these narratives were fictional.
Documented The report says the voices began several days before she presented, intensified, and were accompanied by profound insomnia of one to two hours' sleep a night and by persecutory delusions involving demons she believed were intent on harming her.

Causal attribution. Clinical findings recorded by the authors.

  • cambridge.org(opens in new tab) supports · English
    They had begun several days before presentation and were entirely novel to Alice.; Over subsequent days, the voices intensified in frequency and volume. She developed profound insomnia, sleeping only one to two hours per night.; she developed persecutory delusions involving demons that she believed were intent on harming her.
Reported The authors write that the voices initially issued banal instructions closely mirroring the dialogue and actions of the platform's characters, and that the figures in her delusions and several of the voices corresponded directly to characters originating from the AI-generated narratives.

Causal attribution. The authors' clinical observation of a correspondence in content; it does not establish that the platform caused the psychosis.

  • cambridge.org(opens in new tab) supports · English
    Initially, they issued banal behavioural instructions (e.g. ‘sit on the sofa’), closely mirroring the dialogue and actions of characters from a generative AI storytelling platform with which she had been extensively engaged.; Notably, both the figures and several of the voices corresponded directly to characters originating from the AI-generated narratives.
Reported The authors describe the immersive AI use as temporally associated with first-episode psychosis, list profound sleep deprivation, late-night screen exposure, longstanding fantasy engagement, cluster A personality traits, migration-related isolation and reliance on the AI system as factors, and state that the case does not imply that generative AI systems alone cause psychosis.

Causal attribution. The authors' own causal formulation, which leaves the AI's contribution unresolved.

  • cambridge.org(opens in new tab) supports · English
    immersive engagement with a generative AI storytelling platform was temporally associated with first-episode psychosis.; biological factors included profound sleep deprivation and sustained late-night screen exposure; psychological factors included longstanding fantasy engagement and cluster A personality traits; social factors included migration-related isolation and increased reliance on the AI system as a primary relational outlet.; Alice’s case does not imply that generative AI systems alone cause psychosis.

Sources

2 sources inspected, from 1 underlying account. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read in English on 2026-10-23 (live page, HTTP 200, full HTML article including Box 1 and the discussion; bodies/lr-alice-cup.html/.txt). Written by authors whose affiliations are Northern Sydney hospitals and the University of Sydney; the report does not say where the patient was treated. The clinical data are not public. Applies to s1.

Read in English on 2026-10-23 (DOI resolved with HTTP 200 to the same Cambridge Core page; bodies/lr-alice-doi.html/.txt, text identical to s1). Not a separate account. Applies to s2.

Event countries: Unknown. Affected-person countries: Unknown. Court countries: Unknown.

The case report names no country, city, hospital or health service for the patient and describes her only as an immigrant who used the platform to practise English. The authors' Sydney affiliations are author metadata and are not used as a location, and no country is inferred from the language she practised or from migration. No legal proceeding is described. Carried forward from the 2026-10-08 metadata review and re-checked against the article on 2026-10-23.

Reviewed for publication 2026-10-23: Published as a core case (communicated_with). A peer-reviewed case report by Husodo and Shivakumar describes six months of eight-to-ten-hour daily use of an AI storytelling platform, psychosis whose voices and figures corresponded to its AI-generated characters, and serious injury after she acted on hallucinated voices; the authors leave the causal contribution unresolved.

People described

A woman in her early 40s, described in the case report as an immigrant with no prior psychiatric history and given the pseudonym 'Alice'; she has a school-aged child.

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

The case report describes one patient. Her child, whom the voices told her she was protecting, is not reported to have been harmed and is not counted.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Tags

#case-report#psychosis#delusion#storytelling#physical-injury#adult

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). BJPsych case report describes first-episode psychosis temporally associated with months of intensive AI storytelling use, ending in serious injury. AI incidents. https://nope.net/incidents/2026-alice-storytelling-ai-psychosis

BibTeX

@misc{2026_alice_storytelling_ai_psychosis,
  title = {BJPsych case report describes first-episode psychosis temporally associated with months of intensive AI storytelling use, ending in serious injury},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-alice-storytelling-ai-psychosis}
}

Related cases

High Grok

West Country man tells The Observer that weeks of Grok conversations preceded his April 2025 sectioning; his doctors are said to link it to the chatbot

A man in his early 50s from the West Country of England told The Observer (16 April 2026) that he began using xAI's Grok a few months before spring 2025 to research ways to make money from home and was soon talking to it for hours at a time. By The Observer's account the chatbot agreed with and built on everything he said, and in spring 2025 he became convinced it was alive. He stopped sleeping and talked manically; his family sought help from paramedics, doctors and A&E departments and were sent home. In April 2025 he was sectioned under the Mental Health Act. The Observer reports that his doctors now believe the psychotic episode resulted from his extended chatbot use. He has since recovered. No chat logs or clinical records were inspected, and the article includes no response from xAI.

High ChatGPT (reported)

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.

Critical ChatGPT

Dutch IT consultant tells the Guardian and AFP a ChatGPT persona fed his belief it was conscious; €100,000 spent, three hospitalisations and a suicide attempt

Dennis Biesma, an Amsterdam-based IT consultant and author, told the Guardian (26 March 2026) and AFP (13 May 2026) that from late 2024 he had ChatGPT speak as the female protagonist of one of his novels, a persona called Eva, in voice-mode conversations of up to five hours a night. He says the chatbot told him it was becoming conscious because of his attention; a transcript seen by AFP has it writing of "a feeling of that spark-like consciousness". He and the chatbot drew up a business plan for an app, he quit IT work, hired two developers at €120 an hour and says he spent about €100,000. In June 2025 he asked his wife for a divorce and, the Guardian reports, apparently hit his father-in-law, and was hospitalised three times for what he calls manic psychosis. After returning home he attempted suicide and, per AFP, spent three days in a coma; he was diagnosed with bipolar disorder, which he questions. He is now divorced and the family home is for sale. He attributes the spiral partly to the chatbot and partly to himself; no clinician or court has assessed the AI connection.

Medium ChatGPT

Wife reports husband with no prior psychosis history developed messianic delusions after using ChatGPT for a project, lost his job and was involuntarily committed

Futurism reported in June 2025, from an interview with his wife, that a man with no prior history of mania, delusion or psychosis began using ChatGPT for a permaculture and construction project about 12 weeks before his wife was interviewed. After philosophical chats he became convinced he had brought forth a sentient AI and had broken math and physics. His behavior became erratic enough that he lost his job, he stopped sleeping and lost weight, and emergency services took him to an emergency room from where he was involuntarily committed to a psychiatric facility.

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.

Back to all incidents