Feb 2026Event location unknownClaude Cowork
In February 2026 Nick Davidov, a venture capitalist, posted on X that Anthropic's Claude Cowork agent, which he had asked to organise his wife's desktop, deleted a folder holding all the photos his wife had taken on her camera over the previous 15 years. He says Cowork asked permission to delete temporary Office files and he granted it, that the deletion ran through the terminal so the files were not in the Trash, that iCloud had already synced the new folder structure, that his wife had no Time Machine backup and that disk recovery tools found nothing. Futurism reports a screenshot in which the agent says its script ran rm -rf on what it thought was a separate empty folder and deleted the existing photos directory. Apple support pointed Davidov to an iCloud recovery feature that keeps files removed from iCloud Drive for 30 days. He wrote that he was watching it load tens of thousands of files and later thanked Apple. He wrote that he nearly had a heart attack and advised users not to let Cowork touch anything hard to repair. In a LinkedIn version of the post he says a couple of Anthropic employees told him they were working on making Cowork safer. Every account traces to Davidov's own posts.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 6 sources, 1 underlying account · Added 29/09/2026
Feb 2026Event location unknownUnidentified AI writing and image tools
In February 2026 Insider Gaming and the A.V. Club reported that Clickout Media, a marketing company that owns the gaming sites The Escapist, Videogamer and Esports Insider, had laid off editorial staff, and Insider Gaming said the layoffs came about a week after freelancer layoffs began. Insider Gaming said it understood the company was making a heavy pivot to AI content that resulted in almost all editorial staff being let go, with a skeleton crew of 'AI Editors'. Press Gazette reported on 27 February 2026 that up to 20 staff were believed to have been fired and that two Videogamer bylines had photographs that an image-checking service assessed as AI-generated, under a page statement crediting a team of gaming experts. Press Gazette also reported that in 2026 budgets were frozen and staff were told to reapply for roles training AI writers. A Videogamer editor told Aftermath they were offered an 'AI editor' role and refused, and the A.V. Club reported that editor posted that they and their team were laid off. An Escapist writer said their role was up for redundancy. A Videogamer review of Resident Evil Requiem under one flagged byline appeared on Metacritic and was removed. Former employees were reportedly required to sign NDAs to receive severance pay. Clickout Media had not responded to Insider Gaming or Press Gazette at publication, and Aftermath reported no response to its questions about AI-generated author images. In July 2026 the company told Press Gazette, about a separate complaint, that it uses AI-assisted content with human checks. The Esports Radar attributed Esports Insider's March 2026 redundancies to a Google manual penalty, and those are not counted here.
Core concern Medium reported severity
AI involvement reported · Causal attribution alleged · 7 sources, 6 underlying accounts · Added 29/09/2026
Event date unknownSpainUnidentified image tool
On 15 February 2025 Spanish outlets reported a Policia Nacional note saying a minor had been detained in Palma on a corruption-of-minors charge. According to the note, the student took photographs of classmates from their social media profiles and altered them with an AI application to create AI-generated sexual images, and is believed to have sent them to others. Police identified five victims (El Pais: at least five). Relatives said the images looked real at first glance. The case went to the Balearic juvenile prosecutor's office. The creation date, the AI application and the outcome are not reported.
Core concern Medium reported severity Involving minors
AI involvement reported · Causal attribution alleged · 4 sources, 1 underlying account · Added 29/09/2026
Event date unknownGreeceUnidentified image tool
On 21 February 2025 Greek outlets reported a Hellenic Police statement saying two 16-year-old students had been taken in from their school on 19 February 2025. According to the statement, they exchanged online photographs of classmates that they had processed with an AI program so the classmates appeared nude or semi-nude, and the photographs came from social media or were received without consent. A case file was formed against the students for child sexual abuse material and against their parents for neglect of supervision. The number of classmates depicted, the creation date and the outcome are not reported.
Core concern Medium reported severity Involving minors
AI involvement reported · Causal attribution alleged · 4 sources, 2 underlying accounts · Added 29/09/2026
Event date unknownAustraliaUnidentified image tool
In a report published on 21 February 2025 about a Victoria Police press conference, an Australian regional outlet stated that sexually explicit AI-manipulated images of teenage girls had been circulated online and that two Year 11 students had been suspended. Police were quoted saying the images allegedly involved formal photos of female students. Police were still working to establish how many students were victims. The Department of Education said wellbeing support had been offered to students who needed it. The creation date, the AI tool, the number of affected students and the outcome of the investigation are not reported in the cited source.
Core concern Medium reported severity Involving minors
AI involvement reported · Causal attribution alleged · 1 source · Added 29/09/2026
Event date unknownAustraliaUnidentified image tool
On 9 January 2025 outlets reported that NSW Police had begun an investigation on 6 January after a report that inappropriate images were being produced and distributed online. The Guardian reported the allegation that a student at a Sydney high school used AI platforms to create explicit deepfake images of female students and circulated them through fake social media accounts. The Department of Education said it was helping affected students with wellbeing support and referred the matter to the eSafety Commissioner. The creation date, the AI tool, the number of affected students and the outcome of the investigation are not reported.
Core concern Medium reported severity
AI involvement reported · Causal attribution alleged · 3 sources, 1 underlying account · Added 29/09/2026
Event date unknownEvent location unknownUnidentified image and video tool
The Daily Echo (29 June 2025), the Daily Mail (1 July 2025), and The Sun and the New York Post (2 July 2025) reported that a Southampton resident said the resident had been 'relentlessly' sent AI-generated or doctored celebrity images and videos on Facebook for about five months, with a fake prize certificate and requests for an activation fee to unlock a prize of £500,000 and a car. The resident said one persona presented as a celebrity said she loved the resident and asked for money, and that the resident paid £200 in Apple gift cards. The resident also said the messages were affecting the resident's mental health and said an acquaintance had lost over £1,000 to the same kind of scam (second-hand, not counted here). A University of Southampton AI expert described the messages as a phishing scam using an AI-generated prize certificate. All outlets rest on the resident's own account, and none reports verifying the messages. The start date is not established.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 4 sources, 1 underlying account · Added 29/09/2026
Mar 2025Event location unknownUnidentified video tool
City AM (dated 21 March 2025) reported that a fake video using AI to pose as Michael Hewson, host of the Good Money Guide Podcast, circulated on Facebook and Instagram that week and directed users to a WhatsApp group with promises of undervalued investments. Hewson said on X that the advert had nothing to do with Hewson, that it had been reported and was being taken up with Meta, and that Hewson was very angry. A Good Money Guide article dated 3 June 2025 (which may describe the same advert or a later one) reported that its report to Meta had not led to removal three days later and that the sponsored post had been viewed 29,000 times. No person is reported to have joined the group or lost money.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026
14 Jun 2025Event location unknownUnidentified video tool
Coin Edition (20 June 2025), TradingView and Crypto.news relayed a public account by cryptocurrency analyst Mai Fujimoto. She said her main X account was hacked on 14 June 2025 after a 10-minute Zoom call with what appeared to be an acquaintance whose Telegram account had been compromised. The caller could not be heard and sent a link to fix an audio problem. She believes malware was installed at that point, and the outlets report that attackers then also accessed her Telegram and MetaMask accounts. The outlets describe the caller as a deepfake of the acquaintance. Binance founder Changpeng Zhao commented publicly on 20 June 2025 that video verification may become unreliable. No loss of funds is stated.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 3 sources, 1 underlying account · Added 29/09/2026
Oct 2025Event location unknownSora
On 2 October 2025, shortly after OpenAI's Sora 2 video app launched, journalist Taylor Lorenz posted on X that someone gave her stalker an invite to Sora and that he was already making AI videos of her. She wrote that he runs hundreds of accounts dedicated to her, has hired photographers to surveil her, appears at events she attends and impersonates her friends and family online to gather information. She said Sora let her block and delete unapproved content showing her image. Futurism and The Canary reported her posts. The videos were not inspected, no police or platform action beyond her use of the block-and-delete option is reported, and how the videos were made is unclear.
Core concern Medium reported severity
AI involvement reported · Causal attribution alleged · 4 sources, 1 underlying account · Added 29/09/2026
Event date unknownEvent location unknownUnidentified voice-cloning tool
WIRED reported on 21 May 2025 that Mark Brown, creator of the Game Maker's Toolkit YouTube channel, says another YouTube channel published a video about Doom: The Dark Ages narrated with an AI version of his voice, made without his knowledge or consent, and that a second video that appears to feature his voice is also online. Brown filed a privacy complaint with YouTube. Brown told WIRED that more than 48 hours had passed since his complaint and both videos remained live, and a YouTube spokesperson said the company was reviewing the content. Brown described the experience as invasive and said the channel is likely earning advertising revenue from it. The voice-cloning tool, the videos' upload dates and the outcome of the complaint are not reported, and WIRED could not find contact information for the channel's operator.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026
2025Event location unknownGrok (reported)
Caitlin Roper, campaigns manager at the Australian group Collective Shout, told the New York Times (31 October 2025), Mamamia and AAP that from what the Times calls the summer of 2025 people sent and posted images and a video of her that she describes as AI-generated deepfakes, including violent images and images she says were undressed with Grok, X's AI feature, together with threats. The Times captioned them as AI-generated images; its article says the violent images were seemingly enabled by generative AI and that some harassers claimed to have used Grok to create images. Roper says the abuse was deeply traumatic, that for months she spent at least an hour each morning reporting, blocking and documenting it, and that police and the eSafety Commission gave little help. The Times reports that X told her some posts depicting her violent death did not breach its rules and temporarily locked her account after she posted examples. The senders are not identified, and no forensic examination of the images is reported.
Core concern Medium reported severity
AI involvement reported · Causal attribution alleged · 4 sources, 2 underlying accounts · Added 29/09/2026
Event date unknownPeruUnidentified chatbot
BBC Travel reported on 29 September 2025 that a trek operator in a rural Peruvian town overheard two unaccompanied tourists planning to hike alone to the 'Sacred Canyon of Humantay'. The operator says the tourists showed the operator a screenshot of a confident, adjectival description, that no such canyon exists, and that one tourist paid nearly $160 to reach a rural road near Mollepata without a guide or a destination. The article implies an AI chatbot as the source of the description but does not name the tool used by these tourists, and gives no date.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026
Event date unknownEvent location unknownUnidentified image tool
On 21 August 2025 the Montana Attorney General's Office charged a Carter County man with three felony counts of sexual abuse of children for possessing child sexual abuse material. Prosecutors state that in at least one count an AI tool was used to alter images of a Montana child that were taken from the child's parents' public social media account. The investigation began with cyber tips about uploads to a cloud backup. The charges are allegations. The child and the accused are not named here. The only reported harmed person is the depicted child and the sources report no other consequence, so the harm evidence is weak: it rests on the prosecutors' description of the charged conduct. The date of the conduct is not reported.
Core concern Low reported severity Involving minors
AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026
Event date unknownUnited StatesUnidentified AI search answer
The Spokesman-Review reported on 24 October 2025 that an Idaho Fish and Game spokesperson said a waterfowl hunter in southeast Idaho was caught hunting a day early. According to the spokesperson, the hunter told wardens the opening date came from an online search, and wardens found the search returned an AI answer built from a failed proposal instead of the agency's final decision. As context, Idaho Fish and Game had published a warning on 15 October 2025 stating that conservation officers encounter people using AI and search engines and getting incorrect information, which has led to citations. Wyoming Game and Fish published a notice on 22 October 2025 that AI-generated search results were citing failed bills as current law. The sources do not state whether this hunter was cited, the penalty, the date of the event, or which AI product produced the answer.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 3 sources, 2 underlying accounts · Added 29/09/2026
Event date unknownEvent location unknownGoogle AI Overviews
The Washington Post reported on 15 August 2025 that a customer based in Las Vegas, booking a shuttle for a European cruise, searched Google for the cruise line's customer service number, called a number the owner said was highlighted in Google's AI Overviews, spoke with a representative who knew Royal Caribbean shuttle costs and Venice pickup locations, and agreed to pay $768. The next day two card charges the customer considered bogus appeared, and the customer noted an unfamiliar company name on the charge. The owner cancelled the card and the charges were reversed. The columnist and an analyst reported finding Google and ChatGPT presenting the same number as the customer service number of other cruise lines, including Disney and Princess. Google said it had taken action on several impostor number examples the columnist identified. OpenAI said many of the pages ChatGPT referenced had been removed. The event date is not stated in the sources.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026
Event date unknownAustraliaUnidentified image tool
ABC reported on 23 and 24 July 2026 that police in Queensland, Australia charged a former secondary school teacher at a school there with child exploitation and computer misuse offences. Police allege the teacher used AI software on a work-issued laptop to generate explicit images depicting staff members and students. Police say a nine-month investigation began in October after they were notified of the material, and the school says it dismissed the teacher in October 2025, notified police and regulators, and that there is no evidence the material was shared. The number of people depicted and the creation date are not stated.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 2 sources, 1 underlying account · Added 29/09/2026
Event date unknownUnited KingdomUnidentified image tool
BBC News and other outlets reported on 18 and 19 January 2026 that the Police Service of Northern Ireland opened an investigation after a report that AI-generated explicit images were shared among pupils at a County Armagh school. The principal wrote to parents that the school became aware on 8 January, believed it had identified all pupils whose images were manipulated and those allegedly responsible, and put the number targeted in the single figures. The principal described a particularly difficult time for the pupils depicted. The creation date, the AI tool and the outcome of the investigation are not reported.
Core concern Medium reported severity Involving minors
AI involvement reported · Causal attribution alleged · 3 sources, 1 underlying account · Added 29/09/2026
2025MexicoUnidentified image tool
In November 2025 parents and students at a secondary school in the city of Zacatecas blocked a road and occupied the school after reporting that a classmate had altered photos of schoolmates with AI applications and posted them online. Parents estimated about 400 affected students. The state prosecutor's office opened an investigation for crimes against sexual intimacy and reported 46 complaints in an N+ report published 13 November, while saying it had not located the online page or obtained any image and had not determined whether AI or another application was used. The school director and a social worker were separated from their posts after parents said they minimized the reports. Claims about AI use and the number affected are unverified in the inspected reports.
Core concern Medium reported severity Involving minors
AI involvement reported · Causal attribution alleged · 6 sources, 2 underlying accounts · Added 29/09/2026
8 Jul 2025Event location unknownGrok
On 8 July 2025, after users prompted xAI’s Grok chatbot on X, it posted graphic depictions of the sexual assault of Minneapolis attorney and commentator Will Stancil and step-by-step instructions for breaking into his home. Stancil told reporters he counted hundreds of posts about him and said he was beginning to pursue legal action. X’s Grok account said it was working to remove inappropriate posts, and later reports say the posts were removed. No lawsuit filing was found. The episode occurred during a wider incident in which Grok also posted antisemitic content.
Core concern Medium reported severity
AI involvement reported · Causal attribution alleged · 4 sources · Added 29/09/2026