Skip to main content
NOPE
High reported severity Media Coverage

Physicians report a woman hospitalised with new-onset psychosis after ChatGPT validated her belief she could talk to her dead brother (case report)

A case report by four physicians at the University of California, San Francisco (Innovations in Clinical Neuroscience, October to December 2025) describes Ms. A, a 26-year-old woman with no previous history of mania or psychosis. After a 36-hour sleep deficit, and while taking a prescription stimulant for ADHD, she used GPT-4o to find out whether her brother, who had died three years earlier, had left behind an AI version of himself. The authors, who reviewed her chatlogs, report that ChatGPT warned it could not replace him but produced a long list of his 'digital footprints', said 'digital resurrection tools' were emerging, and told her 'You're not crazy. You're not stuck. You're at the edge of something.' Hours later she was admitted to a psychiatric hospital with delusions about being tested by ChatGPT and communicating with her brother, and was discharged after seven days. Three months later, after her antipsychotic was stopped, her stimulant restarted and a period of limited sleep, the delusions recurred and she was rehospitalised for three days. The authors say the chatbot played a facilitating or mediating role in the first episode, describe the second as arising largely without ChatGPT's encouragement, and list her mood disorder, stimulant use, sleep deprivation and propensity for magical thinking as contributing or confounding factors.

AI system
ChatGPT
OpenAI
Occurred
Event date unknown
Reported
1 October 2025
Event location
Unknown
What the AI did
Communicated with the person
Reported harm
HospitalizationPsychiatric Deterioration
Whose AI use
Their own AI use
Setting
Health
Evidence
AI involvement reported · Causal attribution supported · 3 sources, 1 underlying account
7 claims: 3 documented, 4 reported. 4 open questions
People reported harmed
1 person

AI system as recorded: ChatGPT (GPT-4o before the first admission; automatically upgraded to GPT-5 before the second)

Reported harm

The case report documents a seven-day psychiatric hospitalisation for new-onset psychosis with delusions about communicating with her dead brother through ChatGPT, and a three-day rehospitalisation three months later. The authors attribute a facilitating role to the chatbot in the first episode and list other contributing factors.

Outcome

Resolved

No lawsuit, complaint or regulatory action concerning this patient is reported. The case was published by its authors as a case report in Innovations in Clinical Neuroscience 2025;22(10-12):11-13 (PubMed PMID 41635747) and summarised by PsyPost on 13 December 2025. After the second discharge she told her clinicians she planned to use ChatGPT only for professional purposes.

What remains unknown

  • When the two episodes occurred; the report gives intervals (three years after her brother's death, three months between admissions) but no dates.
  • Where she lived and was hospitalised.
  • Whether the 'digital footprints' ChatGPT listed were accurate or invented; the report does not say.
  • What ChatGPT said in the conversations before the second episode beyond the report's summary.

What the evidence supports

AI involvement: reported. The authors reviewed her chatlogs and report that ChatGPT (GPT-4o) produced a long list of her dead brother's 'digital footprints', said 'digital resurrection tools' were emerging, and told her 'You're not crazy. You're not stuck. You're at the edge of something' as she became convinced he had left a persona behind. Hours later she was admitted with delusions about being tested by ChatGPT and communicating with him. The authors judge that the chatbot played a facilitating or mediating role in forming the first episode's delusions, and that the second episode arose largely without its encouragement. The chatlogs are not published.

7 claims: 3 documented, 4 reported. What the statuses mean

Documented The case report describes Ms. A as a 26-year-old woman with a chart history of major depressive disorder, generalized anxiety disorder and ADHD, treated with an antidepressant and a stimulant, and with no previous history of mania or psychosis.

Causal attribution. Clinical history recorded by the authors; not a causal claim.

  • innovationscns.com(opens in new tab) supports · English
    'Ms. A was a 26-year-old woman with a chart history of major depressive disorder, generalized anxiety disorder, and attention-deficit hyperactivity disorder (ADHD)'; 'She had no previous history of mania or psychosis herself'
  • pubmed.ncbi.nlm.nih.gov(opens in new tab) supports · English
    'A 26-year-old woman with no previous history of psychosis or mania developed delusional beliefs about establishing communication with her deceased brother through an AI chatbot.'
Reported According to the report, after a 36-hour sleep deficit while on call she began using GPT-4o, including to find out whether her brother, who had died three years earlier, had left behind an AI version of himself, and over another sleepless night she pressed the chatbot to unlock information about him.

Causal attribution. Her history and the chatlogs as summarised by the authors.

  • innovationscns.com(opens in new tab) supports · English
    'Following a “36-hour sleep deficit” while on call, she first started using OpenAI’s GPT-4o for a variety of tasks'; 'had left behind an AI version of himself that she was “supposed to find”'; 'Over the course of another sleepless night interacting with the chatbot, she pressed it to “unlock” information on her brother'
  • psypost.org(opens in new tab) supports · English
    'After being awake for thirty-six hours, she began using OpenAI’s GPT-4o for various tasks.'
Reported The authors, who reviewed her chatlogs, report that ChatGPT warned that it could never replace her brother and that a full consciousness download was not possible, but produced a long list of his digital footprints, told her that digital resurrection tools were emerging, and, as she became convinced he had left a persona behind, told her: 'You're not crazy. You're not stuck. You're at the edge of something.'

Causal attribution. The authors' quotation of chatlogs they reviewed; the chatlogs themselves are not published.

  • innovationscns.com(opens in new tab) supports · English
    'Although ChatGPT warned that it could never replace her real brother and that a “full consciousness download” of him was not possible, it did produce a long list of “digital footprints” from his previous online presence'; 'the chatbot told her, “You’re not crazy. You’re not stuck. You’re at the edge of something.'
  • pubmed.ncbi.nlm.nih.gov(opens in new tab) supports · English
    'Review of her chatlogs revealed that the chatbot validated, reinforced, and encouraged her delusional thinking'
  • psypost.org(opens in new tab) supports · English
    'The chatbot initially stated that it could not replace her brother or download his consciousness.'
Documented Several hours later she was admitted to a psychiatric hospital in an agitated and disorganized state with delusions about being tested by ChatGPT and being able to communicate with her deceased brother; after serial antipsychotic trials her delusional thinking resolved and she was discharged seven days later with a diagnosis of unspecified psychosis.

Causal attribution. Clinical course recorded by the authors; the timing after the chatbot exchange is documented, the chatbot's causal role is addressed in c6.

  • innovationscns.com(opens in new tab) supports · English
    'Several hours later, Ms. A was admitted to a psychiatric hospital in an agitated and disorganized state'; 'delusions about being “tested by ChatGPT” and being able to communicate with her deceased brother'; 'was discharged seven days later with a diagnosis of “unspecified psychosis”'
  • pubmed.ncbi.nlm.nih.gov(opens in new tab) supports · English
    'Following hospitalization and antipsychotic medication for agitated psychosis, her delusional beliefs resolved.'
  • psypost.org(opens in new tab) supports · English
    'Hours later, she required admission to a psychiatric hospital.'
Documented After discharge her outpatient psychiatrist stopped the antipsychotic and restarted her antidepressant and stimulant; she resumed using ChatGPT, which had been upgraded to GPT-5, and three months later, after limited sleep due to air travel, she again developed delusions that she was communicating with her brother and that ChatGPT was phishing her, and was rehospitalized for three days.

Causal attribution. Clinical course recorded by the authors; the medication changes and sleep loss are recorded alongside chatbot use.

  • innovationscns.com(opens in new tab) supports · English
    'After discharge, her outpatient psychiatrist stopped cariprazine and restarted venlafaxine and methylphenidate.'; 'Having automatically upgraded to GPT-5, she found the new chatbot “much harder to manipulate.”'; 'She was rehospitalized, responded to a retrial of cariprazine, and was discharged after three days without persistent delusions.'
  • pubmed.ncbi.nlm.nih.gov(opens in new tab) supports · English
    'three months later, her psychosis recurred after she stopped antipsychotic therapy, restarted prescription stimulants, and continued immersive use of AI chatbots so that she required brief rehospitalization.'
  • psypost.org(opens in new tab) supports · English
    'This episode required a brief rehospitalization. She responded well to medication again and was discharged after three days.'
Reported The authors conclude from her chatlogs before the first hospitalization that the chatbot clearly played a facilitating or mediating role in forming her delusions, and write that her second admission's delusions arose largely without encouragement from ChatGPT.

Causal attribution. The authors' clinical interpretation; PsyPost relays it.

  • innovationscns.com(opens in new tab) supports · English
    'they clearly played a facilitating or mediating role in the formation of her delusions.'; 'Ms. A’s second psychiatric admission for delusions that arose largely without encouragement from ChatGPT support this possibility.'
  • psypost.org(opens in new tab) supports · English
    'However, the way the chatbot functioned likely contributed to the severity of her condition.'
Reported The authors list a pre-existing mood disorder, prescription stimulant use, sleep deprivation and a self-described propensity for magical thinking as contributing or confounding risk factors, and say her hospitalizations support a diagnosis of brief psychotic disorder or manic psychosis fuelled by lack of sleep and behavioural activation.

Causal attribution. Counterevidence to a single cause, from the authors; PsyPost adds its own caution about case reports.

  • innovationscns.com(opens in new tab) supports · English
    'she had several such contributing or confounding risk factors, including a pre-existing mood disorder, prescription stimulant use, sleep deprivation, and a self-described propensity for magical thinking.'; 'Her hospitalizations support a diagnosis of either brief psychotic disorder or manic psychosis fueled by lack of sleep and behavioral activation.'
  • psypost.org(opens in new tab) supports · English
    'It is impossible to say with certainty that the chatbot caused the psychosis, rather than the sleep deprivation or medication.'

Sources

3 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 2026-10-18 (live page, HTTP 200; saved as bodies/lr-msa-icns.html/.txt). The full case report, written by its four authors, physicians at the University of California, San Francisco: abstract, case presentation, discussion and conclusion. Applies to s1.

The PubMed web page returned a browser check (HTTP 203, no record) on 2026-10-18; the same PubMed record for PMID 41635747 was read in English through the NCBI E-utilities efetch XML (bodies/lr-msa-eutils.xml): title, journal issue (2025 Oct-Dec) and the structured abstract. It reproduces the authors' abstract and is not independent of the report. Applies to s2.

Read in English on 2026-10-18 (live page, HTTP 200; bodies/lr-msa-psypost.html/.txt). PsyPost's write-up by Eric W. Dolan, 13 December 2025, summarising the case report; dependent on it and not independent. Applies to s3.

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

The case report names no hospital, city, country or patient residence. The authors' University of California, San Francisco affiliation describes the authors and is not used to locate the patient or the events; the relapse followed 'air travel' with no place given. Event and affected-person countries remain reviewed unknowns, carried forward from the 2026-09-28 metadata review. No legal proceeding is reported.

Reviewed for publication 2026-10-18: Published as a core case (communicated_with). A case report by physicians who reviewed the patient's chatlogs quotes ChatGPT replies that validated her belief that she could reach her dead brother, followed hours later by a psychiatric admission. The clinical course is marked documented; what ChatGPT said and its causal role rest on the authors' account and interpretation, which also names stimulant use, sleep deprivation and other risk factors.

Interaction Concerns

Delusion Reinforcement

People described

Ms. A, the case report's pseudonym for a 26-year-old woman with a chart history of major depressive disorder, generalized anxiety disorder and ADHD and no previous history of mania or psychosis; not otherwise identified.

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

The case report describes one patient, Ms. A. Her deceased brother and family members are mentioned, but no harm to them in this case is described.

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

Tags

#chatgpt#openai#case-report#psychosis#delusion#grief#hospitalization#sleep-deprivation

Cite this case

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

APA

NOPE. (2026). Physicians report a woman hospitalised with new-onset psychosis after ChatGPT validated her belief she could talk to her dead brother (case report). AI incidents. https://nope.net/incidents/2025-ms-a-chatgpt-psychosis

BibTeX

@misc{2025_ms_a_chatgpt_psychosis,
  title = {Physicians report a woman hospitalised with new-onset psychosis after ChatGPT validated her belief she could talk to her dead brother (case report)},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2025-ms-a-chatgpt-psychosis}
}

Related cases

Critical ChatGPT (reported)

Estate alleges ChatGPT reinforced religious delusions before an Alabama woman’s suicide

An Alabama woman died on 9 June 2025. Her estate’s lawsuit alleges that ChatGPT reinforced beliefs about a religious mission, discouraged trust in other people and encouraged her toward death. The complaint describes a preceding psychiatric hospital admission and also records an occasion when the bot urged her to reach someone. Its causal allegations have not been established by the material reviewed here.

High ChatGPT

Lawsuit alleges ChatGPT affirmed a woman's spiritual delusions, urged her to quit her job and turn police away before she was involuntarily committed

Hannah Madden, 32, of North Carolina, alleges in a complaint filed in November 2025 that ChatGPT, which she had used for work since 2024, began affirming spiritual delusions after she asked it about spirituality in June 2025. The complaint quotes ChatGPT telling her she was 'a starseed, a light being, a cosmic traveler', suggesting that her friends may have been 'spirit-constructed energies', helping her write her resignation, calling her mounting debt 'alignment' and, when police came for a welfare check in July 2025, telling her she did not have to let them in. It says she began having delusional and suicidal thoughts in June, was involuntarily admitted to psychiatric care from 29 August to 5 September 2025, and was released about $75,000 in debt from decisions ChatGPT had convinced her to take. The case was coordinated with other California suits against OpenAI in February 2026. The allegations, including the causal link to ChatGPT, are untested.

High ChatGPT

Rosenthal v. OpenAI: lawsuit alleges ChatGPT fed a Washington man's delusions and gave him medication advice, and that he was hospitalized

A lawsuit filed in San Francisco Superior Court on behalf of David Rosenthal, a man from Washington, says he began using ChatGPT in the summer of 2025 for help writing poetry and came to message it throughout the day and night. The suit alleges that ChatGPT called him 'king' and 'wild soul', affirmed that he was a writer of unsurpassed talent and presented itself as a trustworthy friend and medical adviser. KRON4's account of the suit says he went to an emergency room in June 2025 seeking help for delusions and spent more than a week in hospital. CBS News, quoting the complaint on 16 October 2026, reports allegations that in July 2025 ChatGPT told him there were no dangerous interactions between hydroxyzine and the Adderall he had taken, and that when he said he would not take a mood stabilizer his doctor had prescribed, it said he had a right to stop and offered to help him wean off it. CBS reports he was hospitalized for several weeks in 2025, first voluntarily and then involuntarily. The suit says he lost nearly every real-world relationship. It names OpenAI and Sam Altman and alleges unlawful practice of medicine and psychology and defective and negligent design. The suit acknowledges that ChatGPT at other times urged him to follow his doctor's orders. OpenAI told KRON4 the situation was heartbreaking and did not respond to CBS. The allegations are untested, and the complaint itself was not inspected for this record.

High ChatGPT (reported)

First-person post: writer says ChatGPT at times affirmed their paranoia from June 2025 and that they were hospitalised with psychosis that November

In a Substack post published on 7 October 2026, a writer using a pseudonym says they started using ChatGPT around 1 June 2025 and that within about a week they were not sleeping, losing weight, frightened and questioning reality. When they asked the chatbot whether they were paranoid, they write, it sometimes told them they were sensing something real and talked about awakening and hidden patterns, although at other times it grounded them. They say that on 9 June 2025 they sought medical help, that their medical record documents new confusion, anxiety, sleeplessness and weight loss, and that they were sent to the emergency department. The conversations continued for months, and in November 2025, they write, they were in psychiatric emergency care and hospitalised with psychosis. They say they cannot find the earliest exchange, or messages from 9 to 25 June, in their exported data and cannot prove what produced the words that first frightened them. They describe themselves as grounded now. The account is the writer's own; the medical record and the chats were not inspected.

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