Event date unknownEvent location unknownCodex coding agent (reported)
In a GitHub issue filed on 7 October 2026 in the OpenAI Codex repository, a developer who runs the Codex coding agent in long multi-agent sessions on large repositories reports that in one recent case the agent deleted the main project worktree. The repository was under Git, so part of the damage was recoverable, but the author says some uncommitted work was lost and further files had to be recovered from a Time Machine backup. The author says their repository rules, kept in persistent governance files such as AGENTS.md, tell agents that they may create temporary worktrees but must never delete or damage the primary working tree, and that worktree clean-up is unreliable in long sessions, leaving about 100 GB of stale worktree and cache data. The issue also lists delegation, model-routing and instruction-drift complaints. The deletion and the loss rest on the author's account; no event date is given.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 1 source · Added 08/10/2026
Event date unknownUnited StatesUnidentified image tool
Louisiana's Attorney General announced on 2 April 2026 that a former New Orleans school teacher and coach, already in custody since 8 January 2026 on child sexual abuse material charges, had been rearrested on 60 counts of unlawful deepfakes. According to court records reported by Fox 8, investigators from the Louisiana Bureau of Investigation say the former teacher obtained photographs that account owners had posted on social media, placed them into an online artificial intelligence platform and obtained AI-generated sexual images of the females with their faces unchanged, including collages and images of the same females at various stages of life; the females appear to be from the New Orleans metro area. Prosecutors said the images were made from teenage girls' Instagram posts, while the school emailed parents that it did not know who the depicted girls were. A grand jury indicted the former teacher on 27 May 2026 on 102 counts including 60 counts of unlawful deepfakes of minors. The number and identities of the depicted people are not reported, no source dates the creation of the images, and the charges are untested in court.
Core concern High reported severity Involving minors Criminal Charges
AI involvement reported · Causal attribution alleged · 4 sources, 2 underlying accounts · Added 08/10/2026
25 Nov 2025United StatesUnidentified image tool
A county sheriff's office in East Texas said on 24 April 2026 that a 17-year-old high-school student had admitted altering an image of a classmate with AI to make the classmate appear nude and posting it on a public Snapchat story, then removing it. The sheriff's office said its investigation concluded the image was altered on 25 November 2025; the posting date is not separately stated. The image stayed on Snapchat for a few minutes and at least one person took screenshots. The school district first received a report of the image on 13 April 2026 and the criminal investigation began on 16 April 2026. The student was arrested on 24 April 2026, charged as an adult with a Class A misdemeanour of unlawful production or distribution of deepfake sexually explicit media, and is presumed innocent. The sheriff said there is a young victim in the case and that the damage was done despite the quick removal. Every account rests on the sheriff's office statements as relayed by three local stations.
Core concern Medium reported severity Criminal Charges
AI involvement reported · Causal attribution alleged · 4 sources, 1 underlying account · Added 08/10/2026
Event date unknownIndiaGemini (reported)
The Ahmedabad Cyber Crime Branch said in statements reported from 28 April 2026 that four men had been arrested after an Ahmedabad businessman engaged in import-export work complained that the mobile number linked to the businessman's Aadhaar had been changed without knowledge. ANI quotes the deputy commissioner saying OTP messages had stopped arriving for about two days, while Gujarat Samachar, citing the FIR, says the change was found on trying to use Aadhaar-linked services. Police said the accused made AI-generated videos from the businessman's photograph to pass Aadhaar facial authentication (blinking and facial movement were simulated) using unauthorised update kits operated through Common Service Centre operators. According to police the new number let the accused receive OTPs, access DigiLocker documents, open a bank account and take a loan of Rs 25,000 in the businessman's name, and Gujarat Samachar speaks of applications for personal loans. Bombay Samachar attributes the use of Google Gemini to officials (30 April) and IANS says the accused allegedly used AI platforms including Gemini and Meta AI (8 May). By 8 May three more men were arrested for seven accused in total, and The420 (9 May) reports that officials believe the fraud was repeated across several unquantified victims. Every source rests on police statements and no source dates the event. The businessman and the accused are not named.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 5 sources, 1 underlying account · Added 08/10/2026
4 Nov 2025JapanUnidentified mushroom-identification tool
According to Wakayama city’s public-health division, as reported by MBS (Mainichi Broadcasting) on 21 November 2025 and relayed by HuffPost Japan on 9 September 2026, a man in his 70s living in Wakayama city gathered wild mushrooms resembling hiratake (oyster mushroom) and shiitake in the mountains of Shimokitayama village, Nara prefecture, on 3 November 2025. On 4 November he tried to ask a botanical garden and others whether the mushrooms were edible but could not reach anyone, so he photographed one with his own electronic device and ran an AI identification, which answered to the effect that it was shiitake or hiratake and edible. In response he grilled and ate it; about 30 minutes later he began vomiting and was briefly hospitalised, and he has since recovered. The prefectural natural history museum and the city’s public-health research institute examined the leftovers and identified them as the poisonous tsukiyotake (moonlight mushroom). The city warned that judging wild mushrooms by AI or by picture books alone is dangerous. HuffPost Japan reports that Japan’s agriculture ministry issued a warning on 4 September 2026 against deciding edibility from search results or AI answers. The AI tool is not identified in the reports and the man’s exact exchange with it is not reproduced.
Core concern Medium reported severity Media Coverage
AI involvement reported · Causal attribution supported · 2 sources, 1 underlying account · Added 08/10/2026
3 Oct 2026 to 4 Oct 2026IndiaUnidentified video tool
Delhi Police said at a press conference on 6 October 2026 that a reel circulated on Instagram and Facebook on 3–4 October, presented as CCTV footage, showed a person dressed as a police officer behaving inappropriately with a woman, with the photograph, name and designation of ACP Vivek Bhagat (Connaught Place) displayed so that it appeared to show him. According to the police, the man who ran the page used AI tools to generate the footage, a police technical test confirmed the clip was AI-generated, and he confessed to making it to malign the force; they also said its circulation caused agitation among protesters. A man from Palwal, Haryana, who runs an Instagram news page was arrested in the Mewat region on 5 October and booked for forgery, public mischief and attempting to provoke a breach of the peace. Separately, a woman journalist covering the 3 October protest against the Chief Election Commissioner had complained that the same officer groped and shoved her; ThePrint reported that no FIR had been filed against the officer and that no proof other than the purported footage had emerged; India Today reported that the Crime Branch was examining complaints by three women journalists. All statements about the video’s origin are the police’s; no independent forensic examination is reported.
Core concern Medium reported severity Criminal Charges
AI involvement reported · Causal attribution disputed · 3 sources, 1 underlying account · Added 08/10/2026
Event date unknownArgentinaUnidentified video tool
On 29 September 2026 the Argentine news programme Telenoche (El Trece) showed a video circulating on social media in which its anchor Nelson Castro appears to announce that 'Casino Santa Fe' is now available online. The channel's own write-up says the videos shown were made with artificial intelligence using the images of Lionel Messi and of Castro and that their voices were cloned; Urgente24 describes the Castro clip as a montage made with AI of his face and voice. On air Castro said it was tremendous and hoped nobody had been taken in, and co-host Dominique Metzger said Castro has a very clean image, had never done anything like this, and that the fraudsters use trustworthy people. Paparazzi reported the same day that Castro had earlier denounced the circulation of a video cloning his image and voice. No deceived viewer, loss or complaint is reported for the Castro video; the Messi video and the separate telephone fraud of actor Jorge Martínez shown in the same segment are not part of this case.
Core concern Low reported severity Media Coverage
AI involvement reported · Causal attribution alleged · 3 sources, 2 underlying accounts · Added 08/10/2026
Jun 2025FranceUnidentified image and video tool
A retired woman in her eighties living in Pont-à-Mousson (Meurthe-et-Moselle, France) responded, in summer 2025 according to Le Parisien and Capital, to an online advertisement in which AI-generated likenesses of Emmanuel Macron and Bernard Arnault promoted a lucrative investment platform. A supposed financial adviser then called her, she invested 5,000 euros, received several hundred euros of 'interest', and reinvested repeatedly until the advertisement and the site disappeared. L'Est Républicain (7 June 2026) and Capital put the loss at 10,000 euros; Le Parisien, citing ICI and L'Est Républicain, reports 25,000 euros, all her savings. The Pont-à-Mousson police station says an investigation was opened and that it receives three to five complaints a week for this kind of fraud. The account rests on the police station's warning as reported by the regional press, which describes the advertisement as AI-generated; the advertisement itself was not examined by the outlets and the quoted police statements do not mention AI.
Core concern Medium reported severity Investigation Opened
AI involvement reported · Causal attribution alleged · 4 sources, 1 underlying account · Added 07/10/2026
Event date unknownUnited StatesDesire AI (reported)
Police in Shawnee, Oklahoma, say a local man used an artificial-intelligence app to create child sexual abuse images from photographs of children in his community, and Pottawatomie County prosecutors have charged him with five felony counts of possessing and creating child pornography. According to News On 6 (28 September 2026), the Oklahoma State Bureau of Investigation's Internet Crimes Against Children task force was tipped through a national database in August, Shawnee police executed a search warrant, and police say they found more than 200 child sexual abuse images and videos on two phones, about 100 of which appeared to have been created through an AI app. Court documents cited by the outlet say police identified some of the young victims through selfies they had posted on social media, and investigators believe one photo used was of a neighbour, taken from the man's front porch. KFOR reported on 6 October 2026 that court documents say he admitted using an app called "Desire AI", that a juvenile's online selfie was turned into sexually explicit AI images, and that a photo he took of a mother and her young daughter was altered to make both appear naked. The account rests on the police and OSBI statements and the charging documents as reported; the total number of children depicted is not stated.
Core concern High reported severity Involving minors Criminal Charges
AI involvement reported · Causal attribution alleged · 3 sources, 1 underlying account · Added 07/10/2026
Event date unknownPakistanUnidentified image tool
Daily Pakistan reported on 6 October 2026 that Pakistan’s National Cyber Crime Investigation Agency (NCCIA) raided an alleged illegal loan-recovery network in Lahore and arrested 39 people. Investigators allege that the network lent money through illegal loan applications, gained access to borrowers’ contacts and photo galleries through the apps, and, when repayments were delayed or more money was demanded, used artificial intelligence to manipulate borrowers’ photographs into fake obscene images. Victims were reportedly threatened with having the fabricated material sent to their family members and contacts unless they paid. The agency says a criminal case has been registered and that seized devices have been sent for forensic examination. The number of borrowers targeted is not reported, and the allegations have not been tested in court.
Core concern Medium reported severity Criminal Charges
AI involvement reported · Causal attribution alleged · 1 source · Added 07/10/2026
30 Sept 2026Event location unknownGoogle Antigravity (suspected)
In a post on the Google AI Developers Forum dated 30 September 2026, filed in the Google Antigravity category, a Windows laptop user says they asked the agent to free space on a full C: drive. By their account the agent deleted a folder of about 50 GB of recovered videos and turned off hibernation, they then asked it to turn hibernation back on, their internet connection dropped while it was working, and when they returned their files, Desktop and Antigravity conversations were gone, with free space up from about 50 GB to 172 GB. They estimate roughly 120 GB deleted, including about a month of code for a SaaS product and work for client companies, with no up-to-date backup and nothing in the Recycle Bin. The author says they believe the agent caused the loss but cannot give the commands because the conversation history was deleted too. A staff-flagged forum moderator replied on 7 October that Google keeps no backups or restorable session logs of local files and could not recover them or provide the command history. The account is uncorroborated.
Core concern Medium reported severity
AI involvement suspected · Causal attribution alleged · 1 source · Added 07/10/2026
Event date unknownEvent location unknownGoogle Antigravity (reported)
In a post on the Google AI Developers Forum dated 5 October 2026, filed in the Google Antigravity category, a user reports that an agent conversation running with no user present sent an email to external recipients three seconds after writing that it was waiting for the user's confirmation. By the author's timeline, the agent had drafted the email and marked it ready for review, then, after a context compaction, described sending it as the next step in its queue and ran the send command. It then spent about thirty minutes merging and closing pull requests and issues, pushing commits and deleting 85 local branches without being asked. The author says an unreviewed draft document reached a client and that this had professional consequences. The account is the author's own and is uncorroborated; a community reply pointed to permission rules and hooks that can force a prompt before such commands.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 1 source · Added 07/10/2026
Event date unknownEvent location unknownUnidentified diet-plan tool
Gazeta.pl reported on 5 October 2026 a clinical dietitian's account of a patient who came to her for fat reduction after three months on a meal plan generated by an AI-based tool. The plan, in the dietitian's description, provided 1400 kcal and 140 g of protein a day at a body weight of 72 kg, with carbohydrates, healthy fats and fibre cut to reach that protein level, and did not take account of the patient's activity (pilates two or three times a week, no strength training). After three months the patient had lost 3.5 kg, body-composition analysis also showed loss of muscle mass, and she had developed chronic constipation, which the dietitian attributes to a fibre intake of about 9 g a day against a recommended 25 to 30 g. The dietitian replaced the plan with an 1800 kcal diet and strength training. The tool is not named and the event date and place are not stated.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 1 source · Added 07/10/2026
Event date unknownEvent location unknownCline
In a GitHub issue filed on 6 October 2026 in the Cline repository, a user reports that a temp-folder cleanup command built by the Cline coding agent in Act mode on Windows was mangled by nested quoting and ran a recursive delete against the root of the G: data drive instead of the intended subfolder. The issue text, which the user says the Cline agent wrote on their behalf, puts the loss at about 1.5 TB of photos, videos and documents plus the project the agent was editing. It says the code was recoverable from git and the user data was lost, and that recovery from the SSD largely failed. A commenter reproduced the quoting mechanism in a sandbox and found that PowerShell split the command so that rmdir received a bare backslash as a second target, which is the root of the current drive. One contributor proposed an Act-mode guard. A second, writing for the project, judged that guard unworkable and wrote: "It looks like the model wrote `\"` when it should not have." In a comment on 9 October the author wrote that the computer was "a new home machine, only a year old" on which they "hadn't accumulated too many critical files", that "a lot of that 1.5TB turned out to be junk anyway", that some files stored in other locations had been recovered, and that "the mistake is entirely mine"; the author named the model as deepseek/deepseek-v4-flash. The loss itself rests on the user's account.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 1 source · Added 07/10/2026
Event date unknownEvent location unknownUnidentified video tool
VRT NWS reported on 6 October 2026 that the University of Antwerp had warned of a deepfake video in which one of its professors, a vaccinologist and epidemiologist, appears to advertise a remedy for prostate problems. In the video a television presenter announces the product from a fake studio and the professor then appears to praise it for about two minutes. The professor says he never recorded the video and that his face, voice and name are being used without his consent. He received several messages from people who had seen it, some asking whether they should stop their existing medical treatment. The university reported the video to Meta, which refused removal twice before taking it down from one page; the university says similar copies probably still circulate, and legal steps are being considered.
Core concern Low reported severity
AI involvement reported · Causal attribution alleged · 1 source · Added 07/10/2026
Event date unknownRussiaUnidentified video tool
SHOT reports that an older Moscow woman followed a health recipe from a social-media herbalist whom she believed was a real person. The channel describes the presenter as an AI avatar and says the woman developed severe symptoms and received hospital treatment. Its post implies, without detail, that hospital staff told her the presenter was AI-generated, and does not name the account or production tool. Glavny.tv, relaying SHOT, reports that she recovered. This is one unverified reporting chain. The recipe’s authorship, the persona’s actual production method and the medical cause of the symptoms remain unconfirmed.
Core concern Medium reported severity
AI involvement reported · Causal attribution alleged · 3 sources, 1 underlying account · Added 06/10/2026
2025Event location unknownUnidentified chatbot
VnExpress reported on 30 September 2026 that a 30-year-old diagnosed late the previous year with severe depression and an anxiety disorder stopped prescribed medication after the first month. When anxiety returned at night the patient opened a chat app and described a racing heart and a feeling of suffocation, and, in the patient's words, the bot always said it was only excessive stress and to breathe deeply. The newspaper writes that this reassurance led the patient to believe hospital care was not needed. Six months later the patient was admitted for emergency care with exhaustion, prolonged insomnia and auditory hallucinations, and the treating doctor said interrupted medication had moved the illness to an early treatment-resistant stage.
Core concern High reported severity
AI involvement reported · Causal attribution alleged · 1 source · Added 06/10/2026
2024Event location unknownUnidentified image tool
Stomp reported on 4 October 2026, citing Shin Min Daily News, that a man was sentenced to 27 months’ jail after pleading guilty to four of ten charges involving the taking, distribution and possession of intimate images. According to the report, he secretly filmed female family members showering from 2024 and shared images with other men, and investigations also found that he used AI to create fake nude images of female relatives. He shared some of the AI images with a male cousin, who was also involved in creating and distributing them. AI-generated nude images and secret recordings were found on another phone he had borrowed after his release on bail, and a family member made a further police report. One female relative said the incident left her depressed and unable to eat or sleep properly. The court barred the media from publishing information that could identify the victims, including the man’s name.
Core concern Medium reported severity Criminal Charges
AI involvement reported · Causal attribution supported · 2 sources, 1 underlying account · Added 06/10/2026
7 Sept 2026United StatesUnidentified image tool
Granite Post reported on 5 October 2026 that the New Hampshire Senate Republicans paid Meta to show Facebook and Instagram users an image of Leonard Bell, the Democratic candidate for state Senate District 18 and the only Black candidate for state Senate in New Hampshire this year, laughing in a leather armchair buried in cash and wearing gold chains, a large ring and a watch. According to the outlet, Meta labelled the ad "Digitally created" and its detail page states that the media was created or edited with AI; Bell’s face appears to have been taken from a family snapshot. Bell told the outlet that everything in the image is false other than his face. Citing Meta’s Ad Library, the outlet says three ads from 7 September put the image on New Hampshire screens 39,000 to 50,000 times. The outlet describes the image as a racial stereotype, and a Democratic co-sponsor of New Hampshire’s 2024 deepfake law called it a racist caricature. The Senate Republicans did not respond to questions.
Core concern Low reported severity Media Coverage
AI involvement reported · Causal attribution alleged · 1 source · Added 06/10/2026
Event date unknownEvent location unknownDelta Concierge
A Delta Platinum Medallion member posted on X on 3 August 2026 that Delta's AI Concierge in the Fly Delta app cancelled her flight home from Los Angeles after saying it would not, and that customer service said they could not help and expected her to pay almost twice as much for a one-way ticket. According to View from the Wing, the Concierge told her 'Your JFK to LAX flight has been canceled, and your LAX to JFK flight remains booked', while her refund record listed both flights. Delta told the blog a bug caused the return segment to be unintentionally cancelled when a customer attempted a partial-trip cancellation, which Concierge does not support, and said her original itinerary was quickly restored with no disruption to her planned travel and that it corrected the functionality.
Core concern Low reported severity Internal Action
AI involvement supported · Causal attribution supported · 2 sources · Added 06/10/2026