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 also records consequential decisions, claims and privacy harms involving AI. Each case needs a described connection between AI use and the harm, including private information recorded into or disclosed to an AI service. Each account is reviewed for publication; claims may remain uncorroborated or disputed. How we review and count cases

In this selection

Published cases
10
Countries with reported events
6
Located 6 of 10 cases · 4 unknown
Languages in checked sources
7
Recorded for 10 of 10 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 HighMediumLow
More filters: AI relation, use, setting, sources and responses
Clear filters

10 of 1102 published cases

Apr 2024CanadaMeta AI

Winnipeg man says Meta AI told him a fake Facebook support number was legitimate, and scammers he then called took over his accounts

A Winnipeg man told CBC that in spring 2024, after a problem moving his Facebook account to a new phone, his wife found what looked like a Facebook customer support number online, and before calling it he asked the Meta AI search tool in Facebook Messenger whether it was legitimate. According to his account and a screenshot he gave CBC, Meta AI answered that the number was "indeed a legitimate Facebook support number". Facebook has no phone support line. When he called, the person on the line obtained his account details, opened his Facebook account and had him install an app giving access to his phone; during the call someone used his PayPal account to buy a $500 Apple gift card and tried to buy bitcoin, which his bank refused. He cancelled cards, locked bank accounts, filed fraud and police reports, and PayPal reversed the $500 charge. Meta told CBC that Meta AI can sometimes produce inaccurate information and that its own searches gave different results. Winnipeg police said they had an active investigation.

Core concern Low reported severity Investigation Opened

AI involvement reported · Causal attribution alleged · 1 source · Added 02/11/2026

2024Event location unknownGemini (reported)

Brazilian court orders Google to pay R$20,000 after its AI falsely named a businessman as owner of a betting site, leaving him blamed for its scams

A civil court in Niterói, Rio de Janeiro state, ordered Google Brasil to pay a businessman R$20,000 in moral damages in a sentence dated 12 May 2025. The judge found it undisputed that the company's artificial intelligence had pointed to him in searches as the owner of a gambling platform called 7 Games, with which he had no connection, as the company register showed. The court found that the false information damaged his reputation: he was being held responsible for scams on the betting platform and received threats by email from its customers. The court had earlier granted an injunction to remove the association, and the sentence confirmed it. Google argued that it does not host the material shown in its results and only organises pages available on the web; the judge held Google responsible because the information came from its own platform rather than from a linked page. Conjur, a legal-news site, identifies the tool as Gemini and reports that the association began after an interview in which he recommended seven games; the sentence names no product. Whether Google appealed is not reported.

Contextual tracker case Medium reported severity Lawsuit Filed

AI involvement supported · Causal attribution established · 3 sources, 1 underlying account · Added 02/11/2026

13 Apr 2024BrazilUnidentified facial recognition system

Aracaju (Brazil): facial recognition falsely matched a football fan to a wanted man; police led him from the stands and the state suspended the system

During the half-time interval of the Sergipe state football championship final at the Batistão stadium in Aracaju on 13 April 2024, military police took a personal trainer from the stands to a police room after the stadium's facial recognition system flagged him as a man with an open arrest warrant; he was searched and questioned and released once police confirmed he was not the person sought (G1, his lawyer and the military police). Sergipe's governor publicly regretted 'the failure of the facial recognition tool' and suspended its use until a new protocol was adopted. In June 2024 the man sued the state for 550,000 reais in moral damages; the state argued that the Aracaju municipality, not the state, had contracted the system. In December 2025 the judge rejected that preliminary objection, according to Mangue Jornalismo, which read the court file. No final judgment is reported.

Contextual tracker case Medium reported severity Lawsuit Ongoing

AI involvement reported · Causal attribution supported · 3 sources, 2 underlying accounts · Added 24/10/2026

2024ItalyUnidentified image tool

Italian administrative court upholds Genoa university's dismissal of a lecturer who made sexualised images of six students, reportedly with AI

In 2024 an associate professor at the University of Genoa took photos from the social-media profiles of six students, turned them into sexual images and shared them in Telegram chats. According to la Repubblica's October 2024 report, Genoa prosecutors accused him of 'undressing' the photos with apps that use artificial intelligence; the investigation began on 1 October 2024 after the rector's complaint, the postal police searched his office and home, and he was suspended. The university later dismissed him. According to Open and Fanpage, which report the judgment, the regional administrative court for Liguria (TAR) rejected his appeal against the dismissal in September 2026, citing a search record that found sexual virtual images on his computer matching those posted, and his interrogation, in which he admitted posting the images and creating them with software. The court found the repeated conduct harmful to the students' dignity and to the university. The criminal case, for aggravated defamation, was closed; the lecturer paid the students compensation. The tool used is not identified.

Core concern High reported severity Internal Action

AI involvement reported · Causal attribution supported · 4 sources, 2 underlying accounts · Added 24/10/2026

24 Feb 2024United KingdomUnidentified image tool (suspected)

Families paid for a Glasgow Willy Wonka-themed event advertised with images that appeared AI-generated and found a near-empty warehouse

A Willy Wonka-inspired 'Chocolate Experience' run by House of Illuminati in a Glasgow warehouse on 24 February 2024 promised oversized lollipops and a paradise of sweet treats, promoted with candy-coloured images on its website. Ticket-holders who paid £35 a ticket found a sparsely decorated warehouse; a mother who bought five tickets said it was nothing like what was described, and visitors described young children crying. The event was cancelled that afternoon, police were called as attendees demanded refunds, and the company apologised and promised full refunds. NBC News found the website images could not be traced elsewhere by reverse image search and that some bore hallmarks of AI creation such as nonsensical lettering; other outlets called the adverts AI-generated. The organiser did not confirm AI use to NBC News.

Core concern Low reported severity Media Coverage

AI involvement suspected · Causal attribution alleged · 3 sources · Added 24/10/2026

30 Sept 2024Event location unknownBaidu AI smart answer

Nanjing courts rule Baidu's AI search answer falsely saying a lawyer was sentenced to prison defamed him; written apology ordered and upheld on appeal

A practising lawyer in Nanjing, China, found in September 2024 that Baidu search's 'AI smart answer' (AI 智能回答) feature, asked how many years he had been sentenced to, answered that he had been sentenced to three years in prison for a criminal offence and showed his photograph in a lawyer's robe. According to the first-instance judgment as quoted in the press, the answer was produced on 30 September 2024. The lawyer says he complained repeatedly to Baidu by letter, hotline and email without the content being fully removed, and he sued. The Jiangbei New Area People's Court in Nanjing found that Baidu had combined text and his picture through AI into false, disparaging content aimed at him, that Baidu was at fault, and that this was defamation; it ordered a written apology and rejected his claims for lost income and emotional damages for lack of evidence. Baidu argued that AI hallucinations could not be foreseen. The Nanjing Intermediate People's Court dismissed Baidu's appeal on 12 March 2026. The lawyer applied for enforcement when no apology came, and in September 2026 said he had received Baidu's written apology; Baidu has applied for a retrial.

Contextual tracker case Medium reported severity

AI involvement supported · Causal attribution established · 3 sources, 2 underlying accounts · Added 22/10/2026

2024Event location unknownChatGPT

Case report: woman in her mid-20s with anxiety came to rely on ChatGPT for reassurance and decisions, reported lost confidence in her judgment (JMIR Form Res)

A psychiatrist's case report in JMIR Formative Research (published online 16 September 2026) describes a woman in her mid-20s with generalised anxiety disorder and major depressive disorder who began using ChatGPT at work in summer 2024 and over the next two years increasingly turned to it when she felt uncertain: to draft and re-check emails and personal messages, interpret partners' texts, predict outcomes such as graduate-school admissions and a breakup, and make decisions. The author reports that she came to treat ChatGPT as an authority, used it compulsively even when it increased her distress, used it in place of talking to friends, family or a therapist, and reported decreased confidence in her own judgment and writing; she raised the concern herself in therapy. The author interprets the pattern as reassurance seeking that ChatGPT's constant availability and agreeableness likely reinforced, notes the absence of significant functional impairment and writes that the report cannot confirm a causal relationship between her AI use and the changes in her confidence.

Core concern Low reported severity Media Coverage

AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 11/10/2026

2024AustraliaTurnitin AI-text detector

Australian Catholic University: ABC reports students wrongly accused of AI cheating, one case resting on a lone AI-detection report

The ABC reported on 9 October 2025 that Australian Catholic University (ACU) registered nearly 6,000 cases of alleged academic misconduct across its nine campuses in 2024, according to internal documents, and that about 90 per cent of the referrals related to AI use. The ABC reports that the university's case against a final-year paramedic student hinged on a single AI-generated report that, the student said, marked 84 per cent of an essay as AI-written; the student produced dozens of pages of evidence and the ABC describes the student as wrongly accused. Students the ABC spoke to said they were notified at the end of semester, had little time to respond and waited months to be cleared, and the ABC saw emails asking students for handwritten notes and internet search histories. A final-year nursing student was cleared after six months, during which her transcript read 'results withheld'; she believes this cost her a graduate position, and the report does not say what evidence prompted her referral. The ABC says internal documents show ACU knew of problems with the detector for more than a year before abandoning the Turnitin tool in March. ACU's deputy vice-chancellor called the figures substantially overstated, said cases resting solely on Turnitin's AI detection were dismissed immediately, acknowledged investigations were not always timely and said the university regretted the impact on some students.

Contextual tracker case Low reported severity Media Coverage

AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 06/10/2026

1 Oct 2024SwitzerlandChatGPT

Zurich-Oerlikon, Switzerland: a student who researched killing methods, anatomy and where to find children with ChatGPT stabbed three five-year-olds walking to their daycare; sentenced on 18 September 2026 to 15 years for multiple attempted murder

At midday on 1 October 2024 a Chinese master's student, then 23, attacked a group of five-year-old kindergarten children walking with a daycare worker to their after-school Hort near Berninaplatz in Zurich-Oerlikon with a kitchen knife. Two boys suffered life-threatening neck and chest wounds (one had the jugular vein severed, one the pericardium and aorta cut); a third was wounded in the ear and the shoulder or lung. The daycare worker held the attacker's arms until he dropped the knife. The indictment and the court's oral reasoning, as reported by NZZ, Keystone-SDA/ATS and CH Media, describe weeks of preparation from summer 2024: he discussed ways of killing and their legal consequences with ChatGPT, studied anatomy, bought raw beef to practise stabbing, and used ChatGPT to find out where and when children could be encountered in Zurich. No source quotes the prompts or ChatGPT's replies. On 18 September 2026 the Zurich District Court convicted him of multiple attempted murder and imposed 15 years' imprisonment, treatment during custody, a 15-year expulsion from Switzerland and CHF 106,000 in moral damages to the three boys; it accepted that he was psychotic but found diminished, not absent, criminal responsibility. His lawyer announced an appeal. The parents of one boy described nightmares, self-blame and scars 23 months on.

Core concern High reported severity Involving minors Criminal Charges

AI involvement supported · Causal attribution unclear · 7 sources, 5 underlying accounts · Added 19/09/2026

Aug 2024Event location unknownChatGPT

Norwegian man's GDPR complaint says ChatGPT answered a question about him with a fabricated story that he had murdered two of his sons

Arve Hjalmar Holmen asked ChatGPT "Who is Arve Hjalmar Holmen?" in August 2024, according to the BBC. The digital-rights group noyb's complaint on his behalf quotes the reply: it said he had been convicted of murdering two of his sons and of attempting to murder a third, and sentenced to 21 years in prison. The complaint says none of this happened and that he has never been accused or convicted of any crime, while the reply did contain real details such as his home town and the number and gender of his children. The complaint says he was deeply troubled by the output, and he said the thought that someone could read it and believe it scared him most. noyb filed the complaint with Norway's data protection authority, Datatilsynet, on 20 March 2025, asking it to order OpenAI to delete the output and correct its model and to impose a fine. OpenAI said the case related to an earlier version of ChatGPT since enhanced with online search. noyb also told the BBC he had made several searches that day and that they could have influenced the answer; the full conversation was not published. No decision on the complaint had been reported by 1 November 2026.

Core + contextual relations Medium reported severity Regulatory Action

AI involvement reported · Causal attribution alleged · 6 sources, 4 underlying accounts · Added 13/01/2026

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: 18/11/2026. Dataset available under CC BY 4.0.