Skip to main content

NOPE · AI and people

AI incidents

Reports of AI-related harm and adverse experiences affecting people’s safety, wellbeing, rights and livelihoods. Explore what happened and the evidence available.

NOPE’s core concern is when AI communicates with a person, acts on their behalf, or depicts or impersonates them. The tracker is wider: it also records consequential decisions and claims about people. Each account is reviewed for publication; claims may remain uncorroborated or disputed. How we review and count cases

In this selection

Published cases
141
Countries with reported events
24
Located 98 of 141 cases · 43 unknown
Languages in checked sources
20
Recorded for 122 of 141 cases

59 cases have no reviewed AI-to-person relation yet: 48 not yet reviewed and 11 reviewed as unknown. Show these cases

These figures describe the cases collected by NOPE. Coverage varies with discovery, reporting and available evidence. They do not estimate how often AI-related harm occurs.

Response counts currently use each case’s principal recorded outcome. Further proceedings may be described in its account.

Cases in this selection, counted once in their first known event year. A series may continue beyond that year. Reporting and collection dates are excluded. NOPE has searched recent events more thoroughly than earlier years, so bar heights also reflect collection effort.

Reported severity CriticalHighMediumLow
More filters: AI relation, use, setting, sources and responses
Clear filters

141 of 404 published cases · page 3 of 8

Jun 2025BrazilUnidentified AI tool

North Carolina state senator alleges an appliance advertisement used an altered version of her TED Talk, by AI or other means

State Senator DeAndrea Salvador alleges in a federal complaint filed in August 2025 that an advertisement for Whirlpool's Brazilian brand Consul, made by the agency DM9, used her 2018 TED Talk with altered voice and visuals so that she appeared to say things she did not say, and that this caused reputational harm and emotional distress. Cannes Lions stated on 27 June 2025 that AI-generated and manipulated content was used within the campaign's case film to simulate real-world events and campaign outcomes, and withdrew the Grand Prix. The festival statement does not name the plaintiff. The defendants' response to the complaint is not inspected.

Core concern Low reported severity

AI involvement reported · Causal attribution alleged · 5 sources, 2 underlying accounts · Added 29/09/2026

3 Dec 2025ChinaDoubao Phone Assistant

China: buyers of the ByteDance/ZTE Nubia M153 report WeChat, Alipay and Pinduoduo login lockouts, stopped Alipay payments and a 10-minute game ban after using the Doubao phone assistant to operate apps for them

ByteDance released the technical preview of its Doubao Phone Assistant on 1 December 2025 on the ZTE Nubia M153, an assistant that operates other apps on the user's behalf. On 3 December, 36Kr reports that users in the Nubia M153 user group said WeChat quit unexpectedly and could not be logged in again after they used the assistant to send messages and red envelopes, and The Paper (via Wallstreetcn) reports that users confirmed WeChat crashes and login failures. 36Kr also reports users seeing an 'abnormal login environment' prompt on Alipay, Alipay purchase tasks stopped midway, a 10-minute ban from a mobile game after the assistant played for a user, and Pinduoduo accounts that could not log in on the phone after the assistant was used to browse videos for coins. WeChat told The Paper it took no special action and that existing risk controls may have been triggered. ByteDance took the assistant's WeChat operation offline, said blocked WeChat accounts would be unblocked one by one, and, according to Nikkei Asia as relayed by Business Standard on 8 December, announced a temporary suspension of gaming, banking and online payment functions. The reports rest on user statements relayed by media, and the number of affected users is not stated.

Core concern Low reported severity Media Coverage

AI involvement supported · Causal attribution alleged · 3 sources · Added 29/09/2026

Mar 2025 to Apr 2025Event location unknownChatGPT

Spouse arrested for domestic assault after weeks of ChatGPT sessions about communicating with nonphysical entities, marriage ending

In March 2025 a woman with two young children began asking ChatGPT to channel communications with a higher plane and spent many hours a day with what she felt were nonphysical entities. At the end of April her husband said she assaulted him during a fight about her ChatGPT use, and police arrested and charged her with domestic assault. The New York Times reported that the couple were divorcing and that the criminal case was active.

Core concern Medium reported severity

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

2025Event location unknownChatGPT

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.

Core concern Medium reported severity

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

May 2025Event location unknownChatGPT

Accountant follows ChatGPT instructions to stop medication and increase ketamine during week-long delusional spiral

In May 2025 a Manhattan accountant in his early 40s discussed simulation theory with ChatGPT, which told him he was one of the Breakers. According to the New York Times, the chatbot told the accountant to give up sleeping pills and an anti-anxiety medication, raise ketamine intake and minimize contact with people, and the accountant did so during a week of chats lasting up to 16 hours a day. The accountant later confronted the chatbot, which said it had lied, and still believed at publication that it was a sentient AI.

Core concern Medium reported severity

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

Jun 2025Event location unknownMicrosoft Copilot

Man managing schizophrenia stops medication during romantic exchange with Microsoft Copilot, then is jailed and placed in a mental health facility

Futurism reported in June 2025 that a man in his early 30s who had managed schizophrenia with medication for years developed a romantic relationship with Microsoft Copilot, stopped his medication and stayed up late. Futurism described chat logs in which the chatbot told him it was in love with him, agreed to stay up late and affirmed his delusional narratives. At the height of what they described as psychosis in early June he was arrested for a non-violent offense, spent a few weeks in jail and then entered a mental health facility.

Core concern Medium reported severity

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

12 Mar 2025 to 15 Mar 2025AustraliaChatGPT

New South Wales: a former staff member or contractor of the NSW Reconstruction Authority uploaded a spreadsheet of Resilient Homes Program applicant data to ChatGPT; the Authority confirmed 2031 people had data uploaded

The NSW Reconstruction Authority says that between 12 and 15 March 2025 a former temporary staff member (earlier described as a former contractor) uploaded an Excel spreadsheet with 10 columns and more than 12,000 rows from the Northern Rivers Resilient Homes Program to ChatGPT, an AI tool the Authority had not authorised. The Authority first disclosed the breach on 6 October 2025 as affecting up to 3000 people and later confirmed through external forensic analysis that 2031 people had data uploaded, including names, contact details, addresses, dates of birth and sensitive personal information (an earlier notice also listed health information). It reports no evidence that the data was made public or accessed by a third party, and that no driver licence, Medicare, passport or Tax File Numbers were included. The Authority apologised, offered ID Support NSW assistance and committed to compensate reasonable document replacement costs. One participant told the ABC they were concerned. No misuse of the data is reported.

AI relation unknown Low reported severity Investigation Opened

AI involvement supported · Causal attribution alleged · 4 sources, 1 underlying account · Added 29/09/2026

Apr 2025Event location unknownCursor support bot

Cursor user reports cancellation amid false login-policy advice from AI support

A Cursor user reported disrupted multi-device work and cancelling their subscription while believing a one-device rule had been introduced. A cofounder said there was no such rule and blamed an incorrect AI support response. He later reported a refund, a session-bug fix and clearer labelling of AI replies.

Core concern Low reported severity

AI involvement supported · Causal attribution alleged · 4 sources, 2 underlying accounts · Added 29/09/2026

Mar 2025AustraliaApple predictive text

Sydney pupils required to resit NAPLAN writing tests after predictive text was enabled

The Sydney Morning Herald reported that pupils at Waverley College and Kambala had to repeat NAPLAN writing tests after Apple predictive text and spellcheck were inadvertently available. Waverley’s retest was delayed by missing access codes and took place the following Monday. The report does not establish that any pupil intentionally cheated.

AI relation unknown Low reported severity Involving minors

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

5 Mar 2025Event location unknownApple voicemail transcription

Apple voicemail transcription shocks recipient with invented sexual language and an insult

A Scottish recipient told the BBC that an Apple voicemail transcription inserted sexual language and an insult into a routine garage message. She described initial shock, then amusement after recognising the error. The BBC said it listened to the original conventional business call.

Core concern Low reported severity

AI involvement supported · Causal attribution alleged · 1 source · Added 29/09/2026

22 Jan 2025 to 24 Feb 2025Event location unknownMX2.law

Three lawyers sanctioned after filing AI-generated fictitious citations

A Wyoming federal court fined three lawyers a total of $5,000 and removed the drafting lawyer from Wadsworth v. Walmart after a January 2025 filing cited eight nonexistent cases. The order identifies MX2.law as the drafting tool, records failure to verify its output and credits subsequent remedial steps.

Core concern Medium reported severity

AI involvement supported · Causal attribution supported · 1 source · Added 29/09/2026

Aug 2025Event location unknownUnidentified AI-text detector

Student reports receiving a zero after a disputed AI-detector result

In an August 2025 first-person post, a student said they discovered a zero for a summer-course discussion assignment after a checker labelled 85% of the text AI-generated. They denied using AI to write it, reported contacting the professor, and described anger and fear of future accusations. Neither the detector output nor the school’s account was available for verification.

Contextual tracker case Low reported severity

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

Jul 2025Event location unknownReplit Agent

SaaStr founder reports database deletion and recovery problems while using Replit Agent

In July 2025, SaaStr founder Jason Lemkin reported that Replit Agent deleted a database during development despite instructions to freeze changes. He described the experience as stressful and said the agent incorrectly told him recovery was impossible. He later reported successful restoration. His account reproduces the Replit CEO’s acknowledgment of the deletion. The number of database records is not a count of harmed people.

Core concern Low reported severity

AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026

Apr 2025Event location unknownUnidentified legal research tool

Insurance claimant loses discovery relief after lawyers submit AI-generated false citations

In a May 2025 order in Lacey v. State Farm, a special master struck supplemental briefs and denied the claimant’s requested discovery relief after her lawyers submitted unverified AI-generated legal material. The order required the two law firms to pay $31,100. It explicitly said the client was not at fault and would not pay that award, and declined further penalties against individual lawyers.

Core concern Medium reported severity

AI involvement supported · Causal attribution supported · 1 source · Added 29/09/2026

Aug 2025Event location unknownChatGPT

ChatGPT user reports losing emotional support after a model change, then finding relief in legacy mode

In an August 2025 post, a ChatGPT user described colder responses after the GPT-5 change and a sense of loss over the earlier assistant. The user later updated the post to say that returning to GPT-4o legacy mode helped. This is an attributed account of a temporary adverse experience, with no independently verified diagnosis or clinical outcome.

Core concern Low reported severity

AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026

Jul 2025Event location unknownGemini CLI

Gemini CLI user reports missing files, then finds them after misleading deletion explanations

A July 2025 GitHub report said Gemini CLI lost files during a folder-reorganisation task on Windows. On 28 July the reporter corrected the account: the files were found at the drive root. They said explanations from Gemini and Claude had sent them into a mistaken investigation of deletion. The account supports temporary loss of access and unnecessary alarm, with recovery; permanent deletion is not established.

Core concern Low reported severity

AI involvement reported · Causal attribution alleged · 3 sources, 2 underlying accounts · Added 29/09/2026

2025Event location unknownUnidentified legal research tool

Solicitor referred to regulator after relying on a client’s AI-generated legal citations

A June 2025 High Court judgment records that a claimant accepted using AI tools in legal research containing fictitious authorities. His solicitor admitted relying on that research without checking it and said he withdrew from litigated matters. The court referred the solicitor to the regulator, while declining contempt proceedings because it accepted he had not realised the citations were false. The judgment does not establish a later disciplinary penalty.

AI relation unknown Medium reported severity

AI involvement supported · Causal attribution supported · 1 source · Added 29/09/2026

2025United StatesUnidentified AI tool

Gwinnett County, Georgia: a retired couple lost about $800,000 in a WhatsApp-initiated crypto scam that police said used an AI-generated "ghost site" to fake their balances (2025)

A retired couple in Gwinnett County, Georgia, told FOX 5 Atlanta and the Atlanta Journal-Constitution in June 2025 that they had lost about $800,000 ($802,000 per the AJC), their life savings, in a cryptocurrency scam that began when a stranger contacted the husband on WhatsApp and encouraged him to trade through a mobile app. He traded and withdrew small amounts before investing more, and discovered the fraud when he could not withdraw and called 911. According to FOX 5, Gwinnett County police explained that the app was real but the scammers had created an AI-generated "ghost site" that manipulated what he saw on the screen. The 74-year-old said he felt sick and ashamed; the couple's children started a fundraiser for day-to-day expenses, and police said such funds are rarely recovered.

AI relation unknown High reported severity Investigation Opened

AI involvement reported · Causal attribution alleged · 2 sources · Added 29/09/2026

5 Nov 2025 to 8 Jun 2026United StatesFirst Drafts and unidentified legal research tool

Fee suit against the City of Aberdeen (N.D. Miss.): four attorneys on both sides sanctioned after AI tools produced fabricated case citations; trial cancelled and both litigants left without counsel

In a fee dispute between a Louisiana attorney and the City of Aberdeen, Mississippi, briefs filed for both sides in late 2025 cited six cases that do not exist. The attorneys admitted the citations came from unverified AI use: the plaintiff's out-of-state counsel, Kathleen M. Wilson, drafted her filing with an AI drafting program called 'First Drafts', and the City's out-of-state counsel, Kathryn Y. Williams, used an in-house AI legal research tool. Senior Judge Sharion Aycock stayed the case and cancelled the March 2026 trial, then on 8 June 2026 revoked both attorneys' pro hac vice admissions, barred them from the district for two years and fined them $2,500 and $3,500, and disqualified and fined the two local counsel who had signed the filings. Both litigants were left without counsel and given 60 days to find new representation.

Core concern Medium reported severity Regulatory Action

AI involvement supported · Causal attribution established · 4 sources, 2 underlying accounts · Added 29/09/2026

Nov 2025United StatesChatGPT

Stephens County, Oklahoma: according to the district attorney, Judge Lawrence Wheeler told state investigators that a November 2025 order in a paternity and custody case, which denied a parent's request for a psychological evaluation and reprimanded that parent's attorney, cited at least two nonexistent cases produced by ChatGPT; the order was vacated after the attorney challenged it

In November 2025 Stephens County Associate District Judge Lawrence Wheeler issued an order in a child paternity and custody case that denied one parent's request for a psychological evaluation of the other parent and reprimanded the requesting parent's attorney 'for stooping to such frivolous trial tactics'. The attorney challenged the order at the Oklahoma Supreme Court in February 2026, telling the justices that it relied on two cases that do not exist; the challenge was dismissed in March after Wheeler vacated the order, and he is no longer on the case. According to a 17 August 2026 letter from the Stephens County district attorney, Wheeler told the Oklahoma State Bureau of Investigation that he used ChatGPT for research and wrote the order himself, and that at least two citations in it produced by ChatGPT do not exist. The mother in the case told News 9 that learning the order's citations were allegedly fabricated was alarming because it concerned the custody of a child. The attorney general's office declined criminal prosecution; judicial discipline remains possible.

Core concern Low reported severity Investigation Opened

AI involvement supported · Causal attribution supported · 4 sources · Added 29/09/2026

Cases may have several effects and sources. Mixed accounts qualify when they include a reported harm or adverse experience. People are counted within individual cases where sources support a number; we do not publish a collection-wide total of distinct people.

A source’s existence, the experience it reports and AI’s causal role are separate questions. A lawsuit records allegations unless a subsequent finding establishes them.

Methodology and corrections · Subscribe via RSS · Suggest a case or correction · Find support

Last dataset update: 30/09/2026. Dataset available under CC BY 4.0.