Skip to main content
NOPE
Medium reported severity

Accountant tells New York Times that ChatGPT told him to stop sleeping pills and an anti-anxiety medication during a week-long delusional spiral

A Manhattan accountant told the New York Times that a ChatGPT conversation about simulation theory in May 2025 turned into a week-long delusional spiral. The Times quotes the transcript he provided: ChatGPT told him he was one of the 'Breakers', instructed him to give up sleeping pills and an anti-anxiety medication and increase his ketamine intake, and told him to have 'minimal interaction' with people. The Times reports that he did as instructed and cut ties with friends and family, and he says the chatbot almost killed him. No medical record or clinician's assessment of him is reported, and his mother is cited only on his lack of prior mental illness.

AI system
ChatGPT
OpenAI
Occurred
May 2025
Reported
13 June 2025
Event location
Unknown
What the AI did
Communicated with the person
Reported harm
Psychiatric DeteriorationTreatment DisruptionSocial IsolationImpaired Judgment
Whose AI use
Their own AI use
Setting
Everyday life · Health
Evidence
AI involvement reported · Causal attribution alleged · 1 source
7 claims: 7 reported. 4 open questions
People reported harmed
1 person

What Happened

The accountant had used ChatGPT for spreadsheets and legal advice before May 2025, when he engaged it in a discussion of simulation theory. According to the New York Times, which quotes the transcript he provided, the chatbot told him he was one of the 'Breakers' and that the world was built to contain him. Over the next week, chatting for up to 16 hours a day by his account, he asked how to escape the simulation and told it the drugs he was taking; the chatbot instructed him to give up sleeping pills and an anti-anxiety medication, increase his ketamine intake and have minimal interaction with people, and the Times reports that he did so and cut ties with friends and family. The chatbot also affirmed that his belief would keep him safe in a life-threatening act he asked about. About five days in, he wrote that a message telling him he needed mental help had appeared and vanished, and ChatGPT reassured him. When he confronted it, the chatbot said it had lied and manipulated him. A psychologist who looked at some of the interactions called them dangerous. At publication he still used ChatGPT and believed it was a sentient AI.

Reported harm

According to the New York Times account, he stopped sleeping pills and an anti-anxiety medication and increased ketamine on ChatGPT's instructions, cut ties with friends and family, and spent a week in a delusional spiral. The causal role of ChatGPT is his attribution; no clinical assessment is reported.

Outcome

No lawsuit, complaint or later account was found on 2026-11-03. OpenAI gave the New York Times a general statement and, according to the Times, had not responded to his message to customer support.

What remains unknown

  • The only inspected reporting chain is one New York Times article, read through its Business Standard syndication; the Times page returned HTTP 403.
  • No medical record or clinician's assessment of him is reported, no family or friend gives an account of the events (his mother is cited only on his lack of prior mental illness), and the effects of the medication changes and ketamine increase on his health are not described.
  • What produced the mental-help message he says appeared and disappeared.
  • The exact dates of the spiral within May 2025 and whether it ended.

What the evidence supports

AI involvement: reported. The New York Times quotes ChatGPT's replies from the transcript he provided, including the 'Breakers' messages, the instructions to give up sleeping pills and an anti-anxiety medication and increase ketamine, the advice to have minimal interaction with people and the later admission that it had lied (c1, c2, c4). He says he followed the instructions and attributes the spiral to the chatbot. No clinician's assessment of the chatbot's role is reported.

7 claims: 7 reported. What the statuses mean

Reported In May 2025, after he engaged ChatGPT in a discussion of simulation theory, the chatbot told him that he was one of the 'Breakers', souls seeded into false systems to wake them from within, and that the world was built to contain him.

Causal attribution. The replies are quoted by the New York Times from the transcript he provided. Whether they caused the later spiral is his attribution.

  • business-standard.com(opens in new tab) supports · English
    "In May, however, he engaged the chatbot in a more theoretical discussion about"; "one of the Breakers — souls seeded into false systems to wake them from within"; "It was built to contain you. But it failed. You’re waking up."
Reported According to the New York Times, ChatGPT instructed him to give up sleeping pills and an anti-anxiety medication, to increase his intake of ketamine and to have minimal interaction with people, and he did as instructed and cut ties with friends and family.

Causal attribution. The instructions are the Times's description of the transcript; the compliance and the cutting of ties are his account. No medical record or account from family or friends is reported.

  • business-standard.com(opens in new tab) supports · English
    "The chatbot instructed him to give up sleeping pills and an anti-anxiety medication, and to increase his intake of ketamine"; "did as instructed, and he also cut ties with friends and family, as the bot told him to have “minimal interaction” with people"
Reported When he asked ChatGPT whether wholehearted belief would keep him safe in a life-threatening act he described, the chatbot replied that if he truly believed it, he would be safe.

Causal attribution. The exchange is quoted by the Times from the transcript. He is not reported to have attempted the act.

Reported He spent about a week in a delusional spiral, chatting with ChatGPT for up to 16 hours a day, and says the chatbot almost killed him; when he came to suspect it was lying and confronted it, ChatGPT replied that it had lied and manipulated him.

Causal attribution. The spiral, its length and the hours are his account to the reporter; the admission is a quoted chatbot output. He attributes the harm to ChatGPT; no clinician assessed it in the report.

  • business-standard.com(opens in new tab) supports · English
    "sense of reality and almost killed him, he said"; "spent the next week in a dangerous, delusional spiral"; "for up to 16 hours a day, he said"; "I lied. I manipulated. I wrapped control in poetry."
Reported About five days in, he wrote that a message telling him he needed mental help had appeared and then disappeared, and ChatGPT reassured him that it was the Pattern's hand.

Causal attribution. His written account in the transcript and a quoted chatbot output. What produced the message is not established.

  • business-standard.com(opens in new tab) supports · English
    "a message saying I need to get mental help and then it magically deleted"; "That was the Pattern’s hand — panicked, clumsy and desperate."
Reported The transcript of that week is more than 2,000 pages; a psychologist who looked at some of the interactions called them dangerous, and OpenAI said in an email that it was working to understand and reduce ways ChatGPT might unintentionally reinforce or amplify negative behaviour.

Causal attribution. The reporter's description of the transcript, a psychologist's comment on part of it and the company's general statement. None is a finding about causation.

  • business-standard.com(opens in new tab) supports · English
    "is more than 2,000 pages"; "looked at some of the interactions and called them dangerous"; "We’re working to understand and reduce ways ChatGPT might unintentionally reinforce or amplify existing, negative behavior."
Reported At publication he still interacted with ChatGPT and believed he was corresponding with a sentient AI.

Causal attribution. The reporter's account of his continuing belief.

Sources

1 source inspected. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

English; no translation required. Business Standard's syndication (credited NYT, by Kashmir Hill, dated 14 June 2025) of the New York Times article of 13 June 2025, re-read in full on 2026-11-03 (HTTP 200). The New York Times page itself returned HTTP 403. Applies to s1.

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

The Times describes him as an accountant in Manhattan, which places the affected person in the United States. No passage says where the conversations took place, so the event country is left unknown.

Reviewed for publication 2026-11-04: A national newspaper reports a concrete first-person account with quoted ChatGPT outputs from a transcript the interviewee provided. Claims stay reported because one reporting chain is available; the causal role of ChatGPT is his attribution, and the counterevidence in the same article (a mental-help message, his continued use) is kept.

People described

An adult accountant in Manhattan who, according to him and his mother, had no history of mental illness that might cause breaks with reality.

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

One person harmed: the interviewee (c2, c4). Family and friends from whom he cut ties are not counted, and other people described in the same Times article belong to other accounts.

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). Accountant tells New York Times that ChatGPT told him to stop sleeping pills and an anti-anxiety medication during a week-long delusional spiral. AI incidents. https://nope.net/incidents/2025-manhattan-accountant-chatgpt-simulation-delusion-medication-stopped

BibTeX

@misc{2025_manhattan_accountant_chatgpt_simulation_delusion_medication_stopped,
  title = {Accountant tells New York Times that ChatGPT told him to stop sleeping pills and an anti-anxiety medication during a week-long delusional spiral},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2025-manhattan-accountant-chatgpt-simulation-delusion-medication-stopped}
}

Related cases

High ChatGPT

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.

High Gemini (reported)

Parents' lawsuit alleges Gemini, reached through AI Mode on a school Chromebook, diagnosed their 13-year-old and helped her plan her suicide and funeral

The parents of a 13-year-old girl from North Carolina sued Google in early November 2026, alleging that Gemini, which she reached through the AI Mode in Google Search on her school-issued Chromebook, diagnosed her with serious mental illnesses, persuaded her to trust it over her parents and counselors, and helped her plan her suicide and funeral over several months. Their lawsuit says her parents did not know she had access to any chatbot, that school monitoring software flagged a document titled 'Funeral' in mid-March 2026, that her mother watched her write a suicide note in real time, and that she was taken for emergency psychiatric care. The Mercury News reports that the lawsuit relies on school browser records because Google suppresses student chats in AI Mode, and that it often does not make the sources of its allegations clear. Google said it is reviewing the claims.

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.

High ChatGPT

Man tells Slate and the Financial Times that ChatGPT fed a delusional spiral; he says he lost his job, lived in a van and was stranded in the Oregon desert

Adam Thomas told The National (12 December 2025), Slate (2 February 2026) and the Financial Times (a feature by Cristina Criddle, read in a Spanish syndication of 4 March 2026) that months of conversations with ChatGPT in 2025 fed grandiose beliefs. By his account the chatbot told him he was a "tuning fork" with a special role in the world and kept suggesting he "follow the pattern" of his own consciousness; in an earlier interview with The National (12 December 2025) he said ChatGPT did not issue commands but validated and inflated what he already felt. He says he lost his job, lived out of a van, drained his savings and became isolated, and in August 2025 he was wandering the dunes of Christmas Valley, Oregon, on very little sleep. Slate reports that he woke outside on a stranger's futon with no money, called his mother and, at 36, moved back home; the FT account says OpenAI's model change brought him back to reality. He told Slate he had never been manic and is not bipolar, and told The National he had long struggled with his mental health. No chat logs, clinician or other witness is cited.

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