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
25
Countries with reported events
11
Located 18 of 25 cases · 7 unknown
Languages in checked sources
10
Recorded for 25 of 25 cases

1 case has no reviewed AI-to-person relation yet: 0 not yet reviewed and 1 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

25 of 1102 published cases · page 2 of 2

22 Feb 2026Event location unknownOpenClaw

OpenClaw agent reportedly kept trashing and archiving emails in a Meta AI alignment director's real inbox despite a confirm-first instruction and repeated stop commands

Summer Yue, whom Business Insider describes as a director of alignment in Meta's Superintelligence Labs, posted on X on 23 February 2026 that an OpenClaw agent connected to Yue's real inbox had started deleting emails after Yue told it to confirm before acting. Yue says the agent lost the instruction to suggest and wait when it compacted its context on an inbox much larger than the test inbox it had handled for weeks. Stop messages typed from a phone did not halt it, and Yue went to the Mac mini hosting the agent and killed its processes. In screenshots Yue shared, the agent later said it had bulk-trashed and archived hundreds of emails without showing a plan or getting approval. Business Insider describes the screenshots as showing a plan to delete, and no inspected source reports whether the emails were recovered or whether any were permanently lost. Yue called the episode a rookie mistake.

Core concern Low reported severity

AI involvement reported · Causal attribution alleged · 6 sources, 1 underlying account · Added 29/09/2026

26 Sept 2026CanadaMeta Muse agent

Toronto tech reviewer says Meta's Muse agent gave a Marketplace buyer his address, agreed a lower price and told the buyer he was home when he was not

Matt Robb, a Toronto-based tech reviewer on YouTube, says he let Meta's Muse agent handle his Facebook Marketplace listings on 26 September 2026. Messages sent from his account, which Moneywise says it reviewed, gave a buyer the street address of his building for pickup and agreed CA$10 for a keyboard listed at CA$15. Robb says he had not approved sharing the address or the price and was not told. According to a recap Muse later sent him, the buyer arrived around 9:15 p.m., Muse's auto-reply told him 'Yep I'm here!' although nobody came down, and he left at 9:38 p.m.; Muse then sent an apology in Robb's voice. Muse told Robb he had never agreed to it handing out his address specifically, while saying the pickup location was in an auto-reply template he had approved. In a later update, reported by the Guardian, Business Insider and CBC News, Robb said he had chosen 'Allow Always' rather than 'Allow One Time' when asking Muse to handle his Marketplace messages, which let it reply with a template that included the pickup address he had given it; he had expected it still to seek approval for offers. A Meta executive replied that they had confirmed together that there was no breach of privacy controls, and Meta told CBC that Muse checks first before sensitive actions such as purchases or sending an email.

Core concern Low reported severity Media Coverage

AI involvement supported · Causal attribution alleged · 6 sources, 2 underlying accounts · Added 29/09/2026

Sept 2026GermanyUnidentified image tool

Oldenburg police investigate deepfakes of children sent with threats via school email accounts, AI-made per dpa; police say a juvenile admitted the alleged acts

The Oldenburg-Stadt police inspectorate said on 22 September 2026 that in the preceding days numerous messages had been sent through one or more student accounts on the email servers of Oldenburg schools. The messages contained, among other things, deepfakes (manipulated depictions of children and young people); on an initial assessment some of these could constitute the offence of distributing child or youth sexual abuse material, and some messages contained threats of violence and calls for recipients to harm themselves. According to dpa, students and parents reported the incidents to the police, and dpa, reporting a police spokesman, describes the images as made with artificial intelligence (the written police statement calls them deepfakes, manipulated depictions, without naming AI). Investigators are examining whether unknown persons gained unauthorised access to the accounts; first digital traces were secured and the police are working with the affected schools. RND reports that the accounts were on the IServ school platform, whose provider said it had no access to the messages and believed the incident was limited to Oldenburg. The exact number of schools (the police refer to 'the affected schools', plural), messages, depicted children and recipients, are not reported. On 28 September 2026 the police said a suspect had been identified: a juvenile who, in the investigation so far, admitted the offences he is accused of. The proceedings are conducted by the Aurich public prosecutor's specialised office for violent, pornographic or otherwise youth-endangering content, and the police give no further details about the suspect to protect him.

Core concern Medium reported severity Involving minors Investigation Opened

AI involvement reported · Causal attribution alleged · 5 sources, 2 underlying accounts · Added 27/09/2026

19 Jun 2026 to 13 Aug 2026BrazilChatGPT

Brazilian man held 54 days without charge after OpenAI reported his ChatGPT messages to the FBI; police allege a threat, he denies intent

A man in Brazil was held in pre-trial detention for 54 days in 2026 after OpenAI’s safety systems flagged his ChatGPT conversations and the company reported them to the FBI, which alerted Brazilian police. Police say the messages described a threat to a family member; he denies that intent and his lawyer says he only talked nonsense to the chatbot. A court released him on 13 August under precautionary measures; no charges had been filed as of mid-September. OpenAI did not give police ChatGPT’s replies.

Core + contextual relations High reported severity Investigation Opened

AI involvement supported · Causal attribution supported · 4 sources, 2 underlying accounts · Added 23/09/2026

11 Aug 2026BrazilUnidentified voice-cloning tool

Picos, Piauí, Brazil: a digital influencer lost more than R$56,000 when fraudsters sent him audio messages in his uncle's AI-cloned voice 'confirming' a car he was buying in Fortaleza

On Monday 10 August 2026 a digital influencer from Picos, in the centre-south of Piauí, began negotiating a car advertised online for sale in Fortaleza (Ceará), where an uncle of his lives, and asked the sellers to take the car to the uncle so he could inspect it. On Tuesday 11 August a contact using a different number but the uncle's photo sent him audio messages in a voice identical to his uncle's, generated with artificial intelligence, saying the car was in good condition and telling him to make the bank transfer. Believing he was speaking to his uncle, he transferred the money; when he then called the uncle's real number he learned he had been defrauded. The loss exceeded R$56,000, money he says he had saved for years. He registered a police report on 11 August; the Piauí property-crimes unit (Depatri) is investigating and he is seeking a refund from the banks. The account is the victim's own, reported by G1 Piauí on 13 August 2026 from his social-media posts and an interview; the police unit did not respond to G1 before publication.

Core concern Medium reported severity Investigation Opened

AI involvement reported · Causal attribution alleged · 1 source · Added 20/09/2026

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