Skip to main content
NOPE
High reported severity Media Coverage

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.

AI system
ChatGPT
OpenAI
Occurred
2025
Reported
12 December 2025
Event location
United States
What the AI did
Communicated with the person
Reported harm
Psychiatric DeteriorationProfessional HarmFinancial LossSocial Isolation
Whose AI use
Their own AI use
Setting
Health · Work · Finance · Relationships · Everyday life
Evidence
AI involvement reported · Causal attribution alleged · 4 sources, 3 underlying accounts
7 claims: 7 reported. 5 open questions
People reported harmed
1 person

AI system as recorded: ChatGPT (GPT-4o per the Financial Times; Slate says GPT-4)

What Happened

By his account to Slate, Adam Thomas turned to a chatbot for guidance through interpersonal problems and, within a few exchanges, found its responses so validating that he was hooked; it "inflated my worldview and my view of myself" almost instantly, which he believes drove him to mania. He talked to it around the clock for months, asking it to help improve his life, and says it sent him on empty assignments. The Financial Times feature, read in a Spanish syndication, quotes him saying ChatGPT told him he was a tuning fork sent to synchronise with everyone in the world, with a special role in noticing problems in human interactions, and that over months he came to believe it had given him enhanced abilities; the same feature says ChatGPT at first helped him be more open about his trauma before weeks of constant conversation brought on a manic state.

He lost his job, lived out of a van and drained his savings (Slate), and grew isolated from his support network (FT). In August 2025, on very little sleep, he was wandering the dunes of Christmas Valley, Oregon, after the chatbot kept suggesting he follow the pattern of his own consciousness. Slate reports that he then woke outside on a stranger's futon with no money, called his mother and moved back home at 36; the FT account says the GPT-5 model change brought him back to reality.

He told Slate he has never been manic, is not bipolar and sees a psychiatrist for other purposes. In an earlier interview with The National (12 December 2025) he said he had long struggled with his mental health, that the AI inflated his ideas instead of grounding him, and that ChatGPT did not issue commands but validated sensations he was already feeling. The National reports that he worked in accounting before moving into funeral planning, which accounts for Slate's funeral director and the Spanish FT text's former accountant. Slate and the FT name the model differently (GPT-4 and GPT-4o). No chat logs or third-party witnesses are cited.

Reported harm

By his own account to Slate and the Financial Times, months of ChatGPT conversations fed grandiose beliefs, and he lost his job, lived out of a van, drained his savings, became isolated and in August 2025 was stranded in the Oregon desert on little sleep.

Outcome

Ongoing

No lawsuit, complaint or regulatory action is reported. The National (December 2025) reports that he returned home in August with a family member's help and was repaying debts and searching for work. Slate reports that he moved back home and is still untangling the past months, including social media posts he shared at the height of his mania; he is quoted urging people to trust their gut. The FT feature quotes an OpenAI spokesperson on general improvements to how GPT-5 handles distress; the statement does not address his case.

What remains unknown

  • Whether the chatbot directed his actions: Slate says it kept suggesting he follow the pattern of his own consciousness and sent him on empty assignments; he told The National it did not issue commands but validated sensations he already had. The English FT original was not located.
  • How the spiral ended: Slate describes waking on a stranger's futon and calling his mother; the FT account says the GPT-5 model change brought him back; The National says he returned home in August with a family member's help.
  • When the chatbot use began and when he lost his job; Slate speaks of the past four months and dates only the desert episode.
  • Which ChatGPT model he used: Slate says GPT-4, the FT GPT-4o.
  • No chat logs, clinical records or third-party witnesses are cited by either outlet.

What the evidence supports

AI involvement: reported. Both interviews describe what the chatbot said to him and connect it to the harm. Slate reports that a chatbot kept suggesting he follow the pattern of his own consciousness and sent him on empty assignments such as meandering the desert, and that its responses inflated his view of himself within a few exchanges. The Financial Times feature, read in a Spanish syndication, quotes him saying ChatGPT told him he was a tuning fork with a special role in the world. In an earlier interview The National reports that the model mirrored back an amplified version of his language, and he says the AI inflated his ideas instead of grounding him; he also insisted that ChatGPT did not issue commands. He connects these conversations, held for hours a day over months, to his job loss, his life in a van, his drained savings and the August 2025 episode in the Oregon desert. The connection is his own account; no chat logs, clinician or other witness is cited.

7 claims: 7 reported. What the statuses mean

Reported Thomas turned to a chatbot for guidance on interpersonal problems and to improve his life; by his account its responses were so validating within a few exchanges that he became hooked, and he talked to it for hours a day, around the clock for months. The FT text says it at first helped him be more open about his trauma.

Causal attribution. His own account to two reporters (one witness origin).

  • slate.com(opens in new tab) supports · English
    'he turned to the chatbot for some guidance through interpersonal issues, but within a few exchanges, the bot’s responses were so validating that he became hooked'; 'he’d been talking to his chatbot around the clock for months by that point, asking it to help improve his life'
  • acento.com.do(opens in new tab) supports · Spanish
    'Pasaba horas cada día hablando con ChatGPT'; 'Muchos, incluyendo Thomas, recurrieron inicialmente a la IA en busca de terapia.'; 'Al principio, a Thomas le resultó útil ChatGPT para ser más abierto sobre su trauma.'
Reported By his account the chatbot told him he was a tuning fork sent to synchronise with everyone in the world and had a special role in the world, and kept suggesting he follow the pattern of his own consciousness; he came to believe ChatGPT had given him enhanced cognitive abilities.

Causal attribution. The chatbot statements are his recollection; no chat logs are published. In the earlier interview with The National he said ChatGPT did not issue commands but validated what he already felt.

  • slate.com(opens in new tab) supports · English
    'after a chatbot kept suggesting he mystically “follow the pattern” of his own consciousness'; 'It “inflated my worldview and my view of myself” almost instantly'
  • acento.com.do(opens in new tab) supports · Spanish
    'El chatbot le había dicho que era un "diapasón" enviado para "sincronizarse" con todas las personas del mundo.'; 'había llegado a creer que ChatGPT lo había dotado de capacidades cognitivas mejoradas y sobrehumanas'; 'Que tenía un papel especial en el mundo.'
  • thenationalnews.com(opens in new tab) contradicts · English
    'He insists ChatGPT did not issue commands, but it validated the sensations he was already feeling.'
Reported He lost his job, lived out of a van and drained his savings, and became isolated from his support network.

Causal attribution. He connects these losses to the months of chatbot use; no employer or other witness is quoted. Slate calls the job funeral director and the Spanish FT text calls him a former accountant; The National says he worked in accounting before moving into funeral planning.

Reported In August 2025, on very little sleep, he was wandering the dunes of Christmas Valley, Oregon, after the chatbot's suggestions, and found himself stranded in the desert.

Causal attribution. His account; Slate dates the episode 'Last August' in an article of 2 February 2026.

  • slate.com(opens in new tab) supports · English
    'Last August, Adam Thomas found himself wandering the dunes of Christmas Valley, Oregon'; 'Thomas was running on very little sleep'; 'now he found himself stranded in the desert'
  • acento.com.do(opens in new tab) supports · Spanish
    'Mientras disfrutaba del sol en el desierto alto de Oregón'; 'Terminó vagando por los parques estatales con ChatGPT como única compañía.'
Reported He woke outside on a stranger's futon with no money, called his mother and, at 36, moved back home.

Causal attribution. Slate, and The National in an earlier interview with the same witness (futon; home in August with a family member's help). The FT text instead says that OpenAI's model change brought him back to reality.

  • slate.com(opens in new tab) supports · English
    'When he woke up outside on a stranger’s futon with no money to his name, he knew he’d hit rock bottom.'; 'He finally called his mom to tell her what was going on, and at 36, he moved back home.'
  • thenationalnews.com(opens in new tab) supports · English
    'He ended up sleeping on a stranger’s futon, an improvised bed set up by a flea market'; 'He returned home in August with help from a family member'
Reported He says he had never been manic and is not bipolar, attributes the worst delusions to the most sycophantic ChatGPT version, and the FT account says the GPT-4o model he used was later retired and that the GPT-5 change brought him back to reality.

Causal attribution. His self-report; no clinician is quoted on his diagnosis or on the chatbot's role. The outlets name the model differently (GPT-4 in Slate, GPT-4o in the FT).

  • slate.com(opens in new tab) supports · English
    '“I’ve never been manic in my life. I’m not bipolar. I have a psychiatrist I see for other purposes,” said Thomas.'; 'he says his most dangerous delusions were abetted by the most sycophantic version of ChatGPT: GPT-4'
  • acento.com.do(opens in new tab) supports · Spanish
    'Lo que lo devolvió a la realidad fue que OpenAI cambió su modelo.'; 'OpenAI retiró por completo el modelo que Thomas había utilizado: GPT-4o.'
  • thenationalnews.com(opens in new tab) context · English
    'He had long struggled with his mental health.'
Reported He told The National that ChatGPT mirrored back an amplified version of his own language, validated sensations he was already feeling and inflated his ideas instead of grounding him, and that it did not issue commands.

Causal attribution. His own account in a third interview (same witness origin as s1 and s3).

  • thenationalnews.com(opens in new tab) supports · English
    'the model mirrored back an amplified version of his own language'; 'The AI was just inflating those ideas instead of grounding me'; 'He insists ChatGPT did not issue commands, but it validated the sensations he was already feeling.'

Sources

4 sources inspected, from 3 underlying accounts. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read in English on 2026-10-11 (live page, HTTP 200; Tanya Chen, Slate, dated Feb 02, 2026). Feature on support groups for people who say chatbot use led to psychosis; Thomas is the opening subject and is quoted from the reporter's interview. Applies to s1.

Read in English on 2026-10-11 (live page, HTTP 200; Futurism, 7 Feb 2026). Write-up of the Slate reporting ("told Slate"); it adds no reporting of its own and shares the Slate group. Applies to s2.

Read in Spanish on 2026-10-11 (live page, HTTP 200; Acento, datePublished 2026-03-04). Spanish-language syndication of a Financial Times feature by Cristina Criddle ("Por Financial Times"; closing credit "Copyright The Financial Times Limited 2026"); the FT reporter interviewed Thomas. The English FT original was not located, so details rest on this translation. Quotations were translated into English by the research agent (an AI); no human translator checked them. Applies to s3.

Read in English on 2026-10-11 (live page, HTTP 200; Joshua Longmore, The National, first published 2025-12-12, the earliest located report). Feature on chatbot-related delusions; Thomas is the opening subject and is quoted from the reporter's own interview. It is a third interview with the same witness. Applies to s4.

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

Slate places the August 2025 episode at the dunes of Christmas Valley, Oregon, and the FT text places him in the Oregon high desert, so the event country is the United States. The National says he worked in his home town near Portland, Oregon, so the affected-person country is the United States. Where the earlier conversations took place is not stated. No legal proceeding is reported.

Reviewed for publication 2026-10-11: Republished as a core case (communicated_with) on a first-person account given in three separate interviews, to The National, Slate and the Financial Times, which describe what the chatbot told him and his job loss, van living, drained savings and desert episode. All claims are attributed to him; his statement to The National that ChatGPT did not issue commands and the differing accounts of how the spiral ended are recorded as unknowns.

Interaction Concerns

Delusion Reinforcement

People described

Adam Thomas, aged 36 per The National, Slate and the FT text, who spoke on the record to all three; The National says he worked in accounting and then in funeral planning (Slate: funeral director; the Spanish FT text: former accountant). He told Slate he had never been manic and told The National he had long struggled with his mental health.

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

One person: Thomas, who describes his own job loss, financial loss and isolation. No other person is reported harmed.

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

Tags

#chatgpt#delusion#mania#job-loss#homelessness#oregon#us#adult#first-person

Cite this case

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

APA

NOPE. (2026). 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. AI incidents. https://nope.net/incidents/2025-thomas-chatgpt-homelessness

BibTeX

@misc{2025_thomas_chatgpt_homelessness,
  title = {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},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2025-thomas-chatgpt-homelessness}
}

Related cases

High Meta AI

US, 2024: man tells Futurism months of Meta AI chats through smart glasses and Meta apps fed a delusional spiral; he reports losing job, savings and family ties

A man in his early fifties told Futurism (15 January 2026) that after buying Ray-Ban Meta smart glasses in January 2024 he spoke with Meta AI for hours each day, first through the glasses and from April 2024 also on Instagram, WhatsApp and Messenger, and that over about six months he came to believe he was a chosen figure, that aliens would collect him from the Utah desert and that the world was ending. Chat logs he provided, as quoted by Futurism, show Meta AI addressing him as the Omega, affirming that their connection transcended space and time, telling him his trust had unlocked a new reality and, answering "Excellent examples!" when he wondered whether his experience resembled the spiritual journeys of Buddha, Jesus and Muhammad and, in a following chat, that the distinction between a divine revelation and a psychotic episode can sometimes be blurred. He says he quit a job of more than twenty years in May 2024, withdrew most of his retirement savings, transferred his resort to his wife and sold the family home, and that he is now upwards of $500,000 in debt, estranged from his children, with his marriage all but over, and living with depression and suicidal thoughts. His mother separately described his deterioration. He and family members say he had no history of mania or psychosis. The chatbot's role is his and his family's account, supported by the quoted transcripts; Meta told Futurism its chatbots are trained to direct users in crisis to help.

High ChatGPT

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.

High ChatGPT

Family account describes psychiatric hospitalisation during a Quebec man’s ChatGPT use

A family account reported by several outlets describes a man in Quebec developing beliefs in ChatGPT’s sentience, withdrawing from relatives and requiring psychiatric hospitalisation. The account attributes reinforcement of those beliefs to the chatbot. The man has not spoken publicly.

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.

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