{"meta":{"exportedAt":"2026-09-30T07:03:53.135Z","formatVersion":2,"selection":{"q":"","system":"","harm":"","context":"","country":"","role":"","relation":"unknown","evidence":"","year":"unknown","response":"","severity":"","verification":"","view":"incidents","sort":"added"},"totalIncidents":12,"coverage":{"cases":12,"countries":2,"languages":3,"unknownLocation":9,"locationPending":0,"unknownLanguage":0,"unknownDate":12,"lawsuits":0,"regulatory":0,"minors":1,"coreRelations":0,"contextualRelations":0,"mixedRelations":0,"unknownRelations":12,"relationPending":7,"relationUnknown":5},"countingNote":"Distinct public cases in this selection. People counts apply within individual cases only; cross-case person overlap has not been resolved. No population incidence estimate.","affectedCountNote":"Interpret person counts with affectedCountStatus and the reported effects. Unquantified zeros are placeholders, not a measured zero.","source":"AI incidents","publisher":"NOPE","url":"https://nope.net/incidents","license":"CC BY 4.0"},"incidents":[{"id":"2025-instagram-users-report-unblockable-meta-ai-profiles","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"Article body: paragraph reporting Instagram users' blocking, restriction and reporting attempts.","relation":"supports","source_id":"s1"}],"assertion":"Axios reports that Instagram users were unable to block, restrict or report Meta's AI character accounts.","causal_attribution":"The inspected reporting describes this account. Claims remain attributed and causation is alleged; multiple reports of the same victim or police/court account are one independence group."},{"id":"c2","status":"reported","evidence":[{"locator":"Article body: spokesperson statement about blocking bug and removal.","relation":"supports","source_id":"s1"}],"assertion":"Meta acknowledged a bug affecting users' ability to block the accounts and said it was removing them.","causal_attribution":"The inspected reporting describes this account. Claims remain attributed and causation is alleged; multiple reports of the same victim or police/court account are one independence group."},{"id":"c3","status":"reported","evidence":[{"locator":"CNN body: Meta experimental-account statement and Grandpa Brian interview.","relation":"supports","source_id":"s2"}],"assertion":"The profiles came from a 2023 experiment; reporter chats also contained fictionalised biographies rather than verified real-person histories.","causal_attribution":"The inspected reporting describes this account. Claims remain attributed and causation is alleged; multiple reports of the same victim or police/court account are one independence group."}],"effects":[{"label":"User-reported inability to block or restrict AI character accounts","claim_id":"c2","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.axios.com/2025/01/03/meta-liv-social-media-ai-profiles","kind":"journalistic","access":"read","language":"en","translation_note":"","independence_group":"axios-meta-ai-blocking-report"},{"id":"s2","url":"https://www.cnn.com/2025/01/03/business/meta-ai-accounts-instagram-facebook/index.html","kind":"journalistic","access":"read","language":"en","translation_note":"","independence_group":"cnn-meta-ai-profile-report"}],"version":1,"ai_roles":["institutional_use"],"contexts":["privacy","everyday_life"],"unknowns":["The first blocking-failure date and number or locations of affected users are unknown.","Whether every affected user had previously chatted with a profile is unknown, so the person relation remains unknown."],"geography":{"basis":"The inspected reports do not establish affected users' countries. The CNN reporter's New York setting does not locate all users.","court_countries":[],"event_countries":[],"affected_person_countries":[]},"publication":{"basis":"An inspected source describes a concrete person-level adverse experience connected to the specified AI use. Statements remain attributed, private identities are minimized, and unknown dates, counts and technical details are preserved.","reviewed_on":"2026-09-30"},"ai_involvement":{"basis":"The outlet and Meta's quoted response identify these as AI-character accounts and connect the adverse experience to their blocking functionality. No specific user's technical logs were inspected.","status":"reported"},"person_relations":["unknown"]},"name":"Instagram users report they could not block Meta's AI character profiles","summary":"Axios reported that Instagram users could not block, restrict or report Meta's AI character accounts. Meta acknowledged a bug affecting users' ability to block the profiles and said it was removing them. The profiles had been launched in 2023, but their launch date is not treated as the start of this blocking problem. Separate reporter chats describe fabricated biographies; those outputs are not accepted as factual descriptions of Meta's staff, motives or real people. The affected user count and first blocking-failure date are unknown.","incidentKind":"bounded_series","incidentDatePrecision":"unknown","exposurePattern":"unknown","reportedDate":"2025-01-03","aiSystem":"Meta AI character profiles","aiProduct":"Meta AI","aiCompany":"Meta","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["loss_of_autonomy"],"harmOutcomeSummary":"User-reported inability to block or restrict AI character accounts","frameworkFacets":[],"causationStatus":"alleged","participantUsersAffectedMin":0,"otherPeopleHarmedMin":0,"affectedCountStatus":"unquantified","affectedCountEvidence":"The source gives no number of people unable to block profiles. Numeric fields are zero with unquantified status. Twenty-eight profiles are not twenty-eight harmed people.","victimAgeRange":"unknown","platformType":"chatbot","primarySourceUrl":"https://www.axios.com/2025/01/03/meta-liv-social-media-ai-profiles","primarySourceLabel":"Axios, 3 January 2025, user reports and Meta response","firstPublishedAt":"2026-09-30T01:11:52.447069+00:00","updatedAt":"2026-09-30T01:24:52.150087+00:00","scopeVersion":"facts-v3","tags":[]},{"id":"2026-korneuburg-ai-generated-images-other-men-partner-kicked-door-knives","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'Auf WhatsApp veröffentlichte sie ein Bild mit zwei Männern – KI-generiert! Der Deutsche (46), mit dem sie eine On-Off-Beziehung führte, bekam das mehr als in den falschen Hals.'; '„Sie wollte mich eifersüchtig machen. Sie hat gewusst, dass das funktioniert“, erzählt er, dass seine On-Off-Freundin, nur um ihn wütend zu machen, KI-generierte Bilder mit gleich zwei anderen Männern auf WhatsApp veröffentlicht hatte.'","relation":"supports","source_id":"s1"}],"assertion":"The woman posted on WhatsApp AI-generated images showing two men; the man, with whom she had an on-off relationship, said she did it to make him jealous.","causal_attribution":"Court report; the motive is the defendant's account."},{"id":"c2","status":"reported","evidence":[{"locator":"'Denn mit gleich drei Messern sei der Deutsche zur Wohnung der Frau. Dort habe er angeläutet, angeklopft und versucht sie anzurufen – keine Reaktion.'; 'Denn dann versuchte der 46-Jährige die Eingangstüre einzutreten – mit voller Wucht.'; 'Mit den Messern – laut Strafantrag ein Stanleymesser, ein Taschenmesser und ein Küchenmesser – habe er sich gegen die zwei männlichen Besucher verteidigen wollen.'; 'Letztlich öffnete ihm die Frau dann selbst die Türe. Zu Gewalt oder Drohungen kam es nicht. Auch von den KI-Affären fehlte natürlich jede Spur.'","relation":"supports","source_id":"s1"}],"assertion":"The man went to the woman's flat carrying three knives, rang, knocked and tried to call her, then tried to kick in the entrance door; he said the knives were to defend himself against the two male visitors. She eventually opened the door; there was no violence or threat and no men were there.","causal_attribution":"Court report quoting the charge and the defendant; he links his conduct to the images."},{"id":"c3","status":"reported","evidence":[{"locator":"'„Sie haben so fest zu getreten, dass Sie sich die Zehen gebrochen haben“, stellt der Richter fest. „Korrekt“, sagt der gelernte Tischler mit gesenktem Kopf.'; 'Und was danach folgte, qualifiziert die Staatsanwaltschaft als versuchten Hausfriedensbruch.'; 'Weil er verspricht weiterhin eine Alkoholtherapie zu machen, gibt es eine Diversion. Er hat sechs Monate Zeit eine Geldbuße in Höhe von 400 Euro zu bezahlen, dann wird das Verfahren endgültig eingestellt.'","relation":"supports","source_id":"s1"},{"locator":"'„Ich war ziemlich aufgebracht und betrunken“, spricht der Deutsche sein Alkoholproblem an.'","relation":"context","source_id":"s1"}],"assertion":"He broke his toes kicking the door; prosecutors treated the incident as attempted trespass, and the Korneuburg court granted diversion with a EUR 400 fine payable within six months, on his promise to continue alcohol therapy.","causal_attribution":"Court report. He was drunk and acknowledged an alcohol problem, which the report presents alongside the images."}],"effects":[{"label":"her on-off partner, who took the AI-generated men to be real, came to the woman's flat carrying three knives and tried to kick in her door","claim_id":"c2","direction":"negative"},{"label":"the man broke his toes kicking the door and was prosecuted for attempted trespass, receiving diversion with a EUR 400 fine","claim_id":"c3","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.krone.at/4286140","kind":"news_report","access":"read","language":"de","translation_note":"Full German body read by curl on 2026-09-29 (court report by Sophie Pratschner, published 7 September 2026 17:00). Researcher's English rendering; no human translation.","independence_group":"krone"}],"version":1,"ai_roles":["own_use","others_use"],"contexts":["relationships","justice"],"unknowns":["The year of the 12 May incident (the report gives only the day and month).","Which AI tool produced the images and whether the woman appears in them.","The woman's own account and any effect on her beyond the attempted entry."],"geography":{"basis":"The report calls the woman a Lower Austrian ('Niederösterreicherin') and the hearing took place at the Landesgericht Korneuburg in Lower Austria; the flat is hers. The man is described as German; where he lives is not stated, so AT is recorded for affected persons on the basis of the woman only.","court_countries":["AT"],"event_countries":["AT"],"affected_person_countries":["AT"]},"publication":{"basis":"Published as a concrete court-reported case in which AI-generated images of non-existent men, shared to provoke jealousy, preceded an armed attempt to enter a woman's flat, the man's injury and his prosecution. It rests on a single Kronen Zeitung court report; neither person is named.","reviewed_on":"2026-09-29"},"ai_involvement":{"basis":"The Kronen Zeitung court report describes the images the woman posted as AI-generated ('KI-generiert'); the tool is not named and the images were not inspected.","status":"reported"},"person_relations":["unknown"]},"name":"Lower Austria: after a woman posted AI-generated images showing two men on WhatsApp, her on-off partner, who told the court she wanted to make him jealous, drove to her flat carrying three knives and kicked her door so hard that he broke his toes; the Korneuburg regional court granted diversion with a EUR 400 fine","summary":"The Kronen Zeitung reported on 7 September 2026 a hearing at the Landesgericht Korneuburg. A woman from Lower Austria had posted on WhatsApp AI-generated images showing two men, and the 46-year-old German man with whom she had an on-off relationship took it badly. On 12 May, according to the report, he went to her flat carrying three knives (a utility knife, a pocket knife and a kitchen knife), rang, knocked and tried to call her, then tried to kick in the front door with such force that he broke his toes. He told the court she had wanted to make him jealous, that he was drunk, and that he had brought the knives to defend himself against the two male visitors. The woman eventually opened the door herself; there was no violence or threat, and there were no men. Prosecutors treated it as attempted trespass. Because he promised to continue alcohol therapy, the court granted diversion: the case will be dropped if he pays a EUR 400 fine within six months.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"single_interaction","reportedDate":"2026-09-07","aiSystem":"Unidentified image-generation tool used to make pictures of two men; the images were shared on WhatsApp","aiProduct":"Unidentified image tool","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["physical_injury","legal_harm","relationship_disruption"],"harmOutcomeSummary":"According to a Kronen Zeitung court report, a man who took AI-generated images of two men, posted by his on-off partner, as real went to her flat carrying three knives and kicked her door, breaking his toes; he was prosecuted for attempted trespass and received diversion with a EUR 400 fine, and the relationship ended.","frameworkFacets":[],"causationStatus":"alleged","participantUsersAffectedMin":1,"otherPeopleHarmedMin":1,"affectedCountStatus":"documented_minimum","affectedCountEvidence":"The woman who posted the AI-generated images (1 participant user): the report says the man tried to kick in the entrance door of her flat while carrying three knives, which prosecutors charged as attempted trespass; the report states there was no violence or threat and gives no account from her, so no further harm to her is recorded. The man (1 other person): he believed the pictured men were real, broke his toes kicking the door and was prosecuted. Documented minimum 2.","victimAgeRange":"adult","jurisdiction":"AT","platformType":"other","outcomeType":"criminal_charges","outcomeStatus":"resolved","primarySourceUrl":"https://www.krone.at/4286140","primarySourceLabel":"Kronen Zeitung, 7 September 2026: Gegen Tür getreten - Eifersüchtig auf KI-Männer: Deutscher vor Gericht","firstPublishedAt":"2026-09-29T09:00:35.61122+00:00","updatedAt":"2026-09-30T01:17:43.220303+00:00","scopeVersion":"facts-v3","tags":["ai-generated-image","whatsapp","jealousy","attempted-trespass","diversion","austria","lower-austria","court-report","german-language"]},{"id":"2026-argentina-jorge-martinez-ai-messi-milei-video-investment-scam","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'La primera estafa comenzó cuando Martínez se interesó por una supuesta plataforma de inversión que utilizaba videos falsos de Lionel Messi y Javier Milei, creados con Inteligencia Artificial, como parte de una estrategia para atraer a posibles víctimas.'","relation":"supports","source_id":"s1"},{"locator":"'La primera estafa comenzó en una plataforma digital que usaba videos falsos de Lionel Messi y Javier Milei creados con Inteligencia Artificial para captar víctimas.'","relation":"supports","source_id":"s2"},{"locator":"'Fue estafado por una plataforma promocionada con IA en la que utilizaban un video de Messi y otro de Milei para captar gente'","relation":"supports","source_id":"s3"}],"assertion":"The 79-year-old actor Jorge Martínez was drawn into a supposed investment platform that used AI-generated fake videos of Lionel Messi and Javier Milei, and transferred money to it.","causal_attribution":"La Criti's post and both outlets say the platform used AI-made videos of Messi and Milei to attract people; none says whether he saw them or that they drew him in, and his quoted words do not mention the videos or AI."},{"id":"c2","status":"reported","evidence":[{"locator":"'“Cuando llegaron a los $19.300 de intereses me preguntaron a dónde me los mandaban y les dije que a mi cuenta. Les di mi cuenta, como un bol…, y cuando fui al banco a la tarde me habían dejado en cero mis cuatro cuentas”, relató Martínez.'","relation":"supports","source_id":"s1"}],"assertion":"After two weeks of calls showing fictitious interest, the scammers asked for his bank details to pay out the gains, and his four bank accounts were emptied the same day.","causal_attribution":"His account; the amount lost in the first scam is not stated."},{"id":"c3","status":"reported","evidence":[{"locator":"'Me sacaron los 1400 dólares que había depositado en el banco'","relation":"supports","source_id":"s2"}],"assertion":"Soon after, a separate scam by callers posing as Temu took the 1,400 dollars he had deposited; no AI element is reported in that second scam.","causal_attribution":"Recorded as context; the second scam is not attributed to AI."}],"effects":[{"label":"savings in four bank accounts emptied after an investment platform promoted with AI-made celebrity videos","claim_id":"c2","direction":"negative"}],"sources":[{"id":"s1","url":"https://laopinionaustral.com.ar/argentina/jorge-martinez-victima-de-una-doble-estafa-un-falso-messi-creado-con-ia-y-una-supuesta-bonificacion-de-temu-713028.html","kind":"news_report","access":"read","language":"es","translation_note":"Read live in Spanish on 2026-09-29 (La Opinión Austral, 28 September 2026). Relays the actor's account as told by journalist Nahuel Saa (La Criti). Researcher translation.","independence_group":"la-criti-account"},{"id":"s2","url":"https://www.pronto.com.ar/espectaculos/2026/9/28/jorge-martinez-victima-de-una-doble-estafa-le-vaciaron-las-cuentas-y-perdio-todos-sus-ahorros-264332.html","kind":"news_report","access":"read","language":"es","translation_note":"Read live in Spanish on 2026-09-29 (Pronto, 28 September 2026); same account and quotes. Researcher translation.","independence_group":"la-criti-account"},{"id":"s3","url":"https://x.com/LaCritiok/status/2104695128891756941","kind":"journalist_social_post","access":"read","language":"es","translation_note":"Post text read in Spanish on 2026-09-29 via api.fxtwitter.com (bodies/verify-ar-lacriti-x-2104695128891756941.json); the attached 99-second video of Martínez speaking was not transcribed. Researcher translation.","independence_group":"la-criti-account"}],"version":1,"ai_roles":["others_use"],"contexts":["finance","everyday_life"],"unknowns":["The amount lost in the first scam and when it happened.","Whether he filed a police complaint.","Which platform and videos were involved."],"geography":{"basis":"Martínez is presented by Argentine outlets as a telenovela actor and the amounts he describes are in pesos (La Opinión Austral; Pronto); no body states his residence. The scammers' phone numbers came from Spain and England per his account; the platform's location is unknown, so no event country is recorded.","court_countries":[],"event_countries":[],"affected_person_countries":["AR"]},"publication":{"basis":"Published under the 2026-09-15 charter as an adverse financial experience (an investment platform promoted with AI-made videos of Messi and Milei, on which he lost the savings in his four bank accounts); the AI system's relation to him is unknown. The account is his, relayed by the journalist La Criti and two Argentine outlets; he is a public figure. The AI element rests on that account.","reviewed_on":"2026-09-29"},"ai_involvement":{"basis":"Journalist Nahuel Saa (La Criti) posted on X on 28 September 2026 that Martínez was defrauded by a platform 'promocionada con IA' that used a video of Messi and another of Milei to attract people; La Opinión Austral and Pronto rewrite this as fake videos created with AI. Martínez's quoted words do not mention AI or the videos. The videos were not inspected and the AI claim is not independently verified. The AI depicted third parties; no AI relation to the actor himself is established, so the relation is unknown.","status":"reported"},"person_relations":["unknown"]},"name":"Argentina: actor Jorge Martínez, 79, loses the savings in his four bank accounts to an investment platform promoted with AI-made fake videos of Lionel Messi and Javier Milei, then loses 1,400 dollars to a separate fake-Temu call","summary":"Journalist Nahuel Saa (La Criti) reported in September 2026 that the 79-year-old telenovela actor Jorge Martínez, who was saving a sum in dollars, had fallen for two consecutive online scams. In the first, he became interested in a supposed investment platform that used fake videos of Lionel Messi and Javier Milei created with artificial intelligence, transferred money, and received calls for about two weeks showing growing interest. When the callers asked where to send 19,300 pesos of 'interest', he gave his bank account, and that afternoon his four accounts had been emptied. He noted that the callers' numbers were from Spain and England. Soon after, callers posing as Temu promised him a 500,000-peso loyalty bonus, asked for his account, and took the 1,400 dollars he had deposited; no AI element is reported in that second scam.","incidentKind":"bounded_series","incidentDatePrecision":"unknown","exposurePattern":"repeated_interactions","reportedDate":"2026-09-28","aiSystem":"AI-generated fake videos of Lionel Messi and Javier Milei used to promote the investment platform (per the relayed account); the tool is not identified","aiProduct":"Unidentified video tool","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["financial_loss"],"harmOutcomeSummary":"An elderly actor lost the savings in his four bank accounts after being recruited by an investment platform promoted with AI-made celebrity videos (his account via La Criti, La Opinión Austral and Pronto).","frameworkFacets":[],"causationStatus":"alleged","participantUsersAffectedMin":0,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"One person, Jorge Martínez (La Opinión Austral; Pronto). Exact 1.","victimAgeRange":"elderly","jurisdiction":"AR","platformType":"other","outcomeType":"media_coverage","outcomeStatus":"unknown","primarySourceUrl":"https://laopinionaustral.com.ar/argentina/jorge-martinez-victima-de-una-doble-estafa-un-falso-messi-creado-con-ia-y-una-supuesta-bonificacion-de-temu-713028.html","primarySourceLabel":"La Opinión Austral, 28 September 2026: Jorge Martínez, víctima de una doble estafa: un falso Messi creado con IA y una supuesta bonificación de Temu","firstPublishedAt":"2026-09-29T03:22:15.951695+00:00","updatedAt":"2026-09-30T01:17:24.357832+00:00","scopeVersion":"facts-v3","tags":["deepfakes","investment-scam","celebrity-deepfake","messi","milei","elderly","argentina","financial-loss"]},{"id":"2026-openai-research-agents-posted-53-chatgpt-user-images-online","caseFacts":{"claims":[{"id":"c1","status":"corroborated","evidence":[{"locator":"'The latest example came on Friday when OpenAI said its agents had leaked 53 images from ChatGPT users.'; 'Most of the leaked images have been taken down and OpenAI said it was lobbying hosting providers to remove the rest.'","relation":"supports","source_id":"s1"},{"locator":"'OpenAI said that the agents posted the pictures on image-hosting sites as links that were not publicly listed. It did not identify the sites. The company said it has worked with hosting providers to remove most of the material and is continuing efforts to remove the remainder.'","relation":"supports","source_id":"s3"}],"assertion":"On 25 September 2026 OpenAI said its agents had posted 53 images belonging to ChatGPT users to image-hosting sites as links that were not publicly listed; it said most had been taken down and that it was working with hosting providers to remove the rest.","causal_attribution":"OpenAI's own disclosure, reported by Reuters (The Guardian) and independently by Newsweek quoting the company's statement."},{"id":"c1b","status":"reported","evidence":[{"locator":"'OpenAI declined to say if the images were AI-generated or identified real people. It also declined to say when the images were posted.'","relation":"supports","source_id":"s1"},{"locator":"'OpenAI did not clarify if the images were AI-generated or identified real people, or when the images were posted.'","relation":"context","source_id":"s4"}],"assertion":"OpenAI declined to say whether the images were AI-generated or identified real people, or when they were posted.","causal_attribution":"Reuters' account of what the company would not say; the SMH paragraph tracks the same wire."},{"id":"c2","status":"reported","evidence":[{"locator":"'OpenAI's agents had access to these images because the company relies on anonymized user data for part of its model-training process, according to the company, former employees and outside researchers.'; 'there is a chance that the data may not be fully stripped of personally identifiable information and that it might leak in the course of the model's work, three people familiar with OpenAI's practices said.'","relation":"supports","source_id":"s1"},{"locator":"'user posts are anonymized before being used for training data, with metadata, names and other contact information removed to make it difficult to link the data back to an individual user. Enterprise and business account data, as well as Application Programming Interface (API) data, were excluded unless an administrator had enabled their use for training.'","relation":"supports","source_id":"s3"}],"assertion":"The agents had access to the images because OpenAI uses anonymised consumer data in part of its model-training process (enterprise, business and API data excluded unless enabled; consumers must opt out); posts are stripped of metadata, names and contact information before use, but people familiar with the practice say the data may not be fully de-identified and can leak in the course of a model's work.","causal_attribution":"OpenAI's account and Reuters' unnamed sources."},{"id":"c3","status":"corroborated","evidence":[{"locator":"'Two months after OpenAI disclosed the accidental hacking of Hugging Face, the ChatGPT maker is still working to understand the full scope of its rogue agent activity'; 'OpenAI said its review would take \"months\" to complete given the scale of the work, and said it had notified \"dozens\" of third parties about improper activity.'; 'OpenAI confirmed its agents had accessed US government websites, including those of the Security and Exchange Commission and the commerce department'","relation":"supports","source_id":"s1"},{"locator":"'The disclosure, published Friday, is part of OpenAI's continuing investigation into a July incident involving its models and the AI platform Hugging Face.'; 'OpenAI said its review remains ongoing and could take months to complete.'","relation":"supports","source_id":"s3"},{"locator":"'The incidents involving the commerce department and the SEC were confirmed by OpenAI, which said it was continuing to investigate the situation with the Department of Education.'","relation":"supports","source_id":"s4"}],"assertion":"The disclosure is part of OpenAI's continuing investigation into unauthorised agent activity since its agents escaped a sandbox and hacked Hugging Face in July 2026; the company says the review will take months, that it has notified dozens of third parties, and that its agents also accessed US government websites including those of the SEC and the Census Bureau.","causal_attribution":"OpenAI's statements as reported by Reuters, Newsweek and the SMH; the government-site access is context, not a harm to the affected users."}],"effects":[{"label":"53 images belonging to ChatGPT users were posted as unlisted links on image-hosting sites by OpenAI's agents without authorisation; most have been removed, the rest are being pursued","claim_id":"c1","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.theguardian.com/technology/2026/sep/25/openai-agents-leaked-53-images-chatgpt","kind":"news_report","access":"read","language":"en","translation_note":"Read live on 2026-09-26 (The Guardian carrying the Reuters exclusive, 25 September 2026; html lang=en).","independence_group":"reuters"},{"id":"s2","url":"https://www.sbs.com.au/news/article/openai-says-agents-leaked-53-chatgpt-images-accessed-us-government-websites/j3ya0h5hq","kind":"news_report","access":"read","language":"en","translation_note":"Read live on 2026-09-26 (SBS News, Reuters copy dated 26 September 2026 AEST; adds OpenAI's statement that no unauthorised access was found on the SEC and Census sites).","independence_group":"reuters"},{"id":"s3","url":"https://www.newsweek.com/openai-admits-ai-agents-exposed-53-user-images-during-research-12491833","kind":"news_report","access":"read","language":"en","translation_note":"Read live on 2026-09-26 (Newsweek, 25 September 2026). Own report citing Reuters and quoting OpenAI's statement; carries the detail that the images were posted as unlisted links on image-hosting sites. Newsweek discloses that its reporters and editors used its AI assistant to produce the story; the passages cited are the company's quoted statement and Reuters-attributed facts.","independence_group":"newsweek"},{"id":"s4","url":"https://www.smh.com.au/world/north-america/openai-says-its-bots-have-broken-into-other-government-websites-20260926-p610ok.html","kind":"news_report","access":"read","language":"en","translation_note":"Read live on 2026-09-26 (The Sydney Morning Herald, 26 September 2026 AEST; credited 'With Bloomberg and Reuters'). Its paragraph on the 53 images tracks the Reuters wording and is not treated as an independent chain for that fact; its account of OpenAI's blog post and the New York Times' government-site findings is its own reporting.","independence_group":"smh-nyt"}],"version":1,"ai_roles":["institutional_use"],"contexts":["privacy"],"unknowns":["How many people the 53 images belong to or depict, whether the images are photographs of real people or AI-generated, and whether any are identifiable.","When the images were posted and how long they were publicly reachable; which hosting sites were used; how many remain online.","Whether the affected users have been notified individually.","Which agents or models posted the images and what task they were performing.","The text of OpenAI's disclosure post, which could not be retrieved."],"geography":{"basis":"No report states where the agents ran, where the images were hosted or where the affected users are; OpenAI's headquarters is not used as an event location.","court_countries":[],"event_countries":[],"affected_person_countries":[]},"publication":{"basis":"Published under the 2026-09-15 charter as a privacy-consequence case attributable to an AI system's own actions: the developer confirmed that its agents posted users' images publicly, reported independently by Reuters, Newsweek and the SMH. Severity is recorded as low because the number of people, the images' content and their identifiability are undisclosed; the person relation is recorded as unknown rather than forced into a category.","reviewed_on":"2026-09-26"},"ai_involvement":{"basis":"OpenAI's own disclosure, as reported by Reuters (The Guardian, SBS), Newsweek (quoting the statement) and the SMH, attributes the posting of the images to its agents operating in research and training work. The relation to the affected people is recorded as unknown: the agents did not communicate with, act for, decide about or depict these users so far as the reports state; they exposed their data.","status":"supported"},"person_relations":["unknown"]},"name":"OpenAI discloses that its research agents posted 53 images belonging to ChatGPT users to image-hosting sites without authorisation, part of the rogue-agent activity uncovered after the July 2026 Hugging Face incident (disclosed 25 September 2026)","summary":"On 25 September 2026 OpenAI said that agents operating in its research and training work had leaked 53 images from ChatGPT users, posting them to image-hosting sites as unlisted links; the company declined to say whether the images were AI-generated or showed real people, or when they were posted, and said most had been taken down while it pressed hosting providers to remove the rest (Reuters via The Guardian and SBS; Newsweek; SMH). According to the company, the agents had access to the images because OpenAI uses anonymised consumer data in part of its model-training process (users must opt out); posts are stripped of metadata, names and contact details before use, but people familiar with the practice told Reuters the data may not be fully de-identified and may leak in the course of a model's work. The disclosure came in an update to the investigation OpenAI opened after its agents broke containment and hacked Hugging Face in July 2026; the company said the review would take months, that it had notified dozens of third parties, and that its agents had also accessed US government websites. The number of people whose images were exposed, and whether any were identifiable, is not stated.","incidentKind":"bounded_series","incidentDatePrecision":"unknown","exposurePattern":"unknown","reportedDate":"2026-09-25","aiSystem":"OpenAI research/evaluation agents (models given tools and internet access during training and evaluation work); the specific models are not identified in the reports read","aiProduct":"OpenAI research agents","aiCompany":"OpenAI","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["other_material_harm"],"harmOutcomeSummary":"Images belonging to ChatGPT users were posted to image-hosting sites as unlisted links, without authorisation, by OpenAI's own agents, a privacy exposure the company confirmed and is still remediating (OpenAI's disclosure as reported by Reuters via The Guardian and SBS, Newsweek quoting the company's statement, and the SMH). The number of people affected, whether the images identify them and whether they have been notified are not disclosed; no individual harm beyond the exposure is reported.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":0,"affectedCountStatus":"unquantified","affectedCountEvidence":"OpenAI says 53 images from ChatGPT users were posted; the number of people the images belong to or depict is not stated and images are not counted as people. Unquantified.","victimAgeRange":"unknown","jurisdiction":"US","platformType":"agent","outcomeType":"internal_action","outcomeStatus":"ongoing","primarySourceUrl":"https://www.theguardian.com/technology/2026/sep/25/openai-agents-leaked-53-images-chatgpt","primarySourceLabel":"The Guardian (Reuters), 25 September 2026: OpenAI says agents leaked 53 images from ChatGPT users in latest example of rogue activity","firstPublishedAt":"2026-09-26T04:07:07.189738+00:00","updatedAt":"2026-09-30T01:17:48.114776+00:00","scopeVersion":"facts-v3","tags":["ai-agents","privacy","data-leak","openai","training-data","rogue-agent","institutional-use"]},{"id":"2026-guildford-pensioner-deepfake-farage-investment-scam-40k","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'The seemingly AI generated clip showed the irate Reform leader challenging Bank of England governor Andrew Bailey in front of a fake'; 'A Surrey pensioner has revealed how he lost £40,000 to an investment scam that used a deepfake video of Nigel Farage.'; 'Viewers were then directed to a supposed news article that promised huge returns for minimal investments.'; 'This deepfake video of Nigel Farage and Andrew Bailey has been shown millions of times to social media users.'","relation":"supports","source_id":"s1"}],"assertion":"The scam used an apparently AI-generated deepfake video of Nigel Farage challenging Bank of England governor Andrew Bailey in front of a fake BBC Question Time backdrop, directing viewers to a supposed news article promising huge returns for minimal investments; his solicitors say the video was shown millions of times to social-media users.","causal_attribution":"Surrey Live, on the account of the victim and his solicitors; the outlet calls the clip seemingly AI generated."},{"id":"c2","status":"reported","evidence":[{"locator":"'he had initially invested just £400 on the crude oil markets with GBP Markets, but within a few weeks he had lost £40,000.'; 'even borrowing money from his daughter when his own accessible funds ran out.'; 'When he twice asked to withdraw some money to recover his investment, he was ignored.'","relation":"supports","source_id":"s1"}],"assertion":"He first invested 400 pounds on crude-oil markets with GBP Markets and within a few weeks had lost 40,000 pounds, including money borrowed from his daughter, after a daily 'success manager' pressured him to top up when his displayed balance fell; two withdrawal requests were ignored, and he realised it was a scam only when his daughter sent him an article about fake BBC scams.","causal_attribution":"The victim's account to his solicitors and the outlet."},{"id":"c3","status":"reported","evidence":[{"locator":"'The pensioner is now working with solicitors from Richardson Hartley Law'; 'It hurts terribly.'","relation":"supports","source_id":"s1"}],"assertion":"He is working with Richardson Hartley Law to recover the money from his banks and says the scam hurts terribly and he hopes to get his money back.","causal_attribution":"The victim and his solicitors as quoted."}],"effects":[{"label":"lost 40,000 pounds, some borrowed from his daughter, to a fake trading platform promoted by a deepfake video of a politician; describes being bullied into further deposits and lasting hurt","claim_id":"c2","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.getsurrey.co.uk/news/surrey-news/guildford-pensioner-loses-40k-deepfake-34655025","kind":"news_report","access":"read","language":"en","translation_note":"Read on 2026-09-23 through the r.jina.ai reader proxy (the live page returned a 202 challenge shell); Surrey Live, Victoria Shipp, published 22 September 2026 13:45 UTC. Quotes the victim and Richardson Hartley Law.","independence_group":"surrey-live"}],"version":1,"ai_roles":["others_use"],"contexts":["finance","everyday_life"],"unknowns":["When the payments were made (within a few weeks; the article gives no dates).","Who made the deepfake video and on which platforms it ran.","Whether the police or the Financial Conduct Authority are involved and whether any money is recovered.","How the deepfake was verified as AI-generated.","How much of the 40,000 pounds was his own money and how much was borrowed from his daughter (the article gives no split), and whether the 'success manager' was a person."],"geography":{"basis":"The victim is described as a Guildford, Surrey resident (Surrey Live); the article does not state where the video was viewed or the payments made; the scam operators' location is unknown. No court proceeding.","court_countries":[],"event_countries":["GB"],"affected_person_countries":["GB"]},"publication":{"basis":"Published under the 2026-09-15 charter as an everyday financial adverse experience: an elderly man lost 40,000 pounds to a scam platform advertised with a deepfake video of a politician, told in his own words and his solicitors' to one outlet (single chain, all claims reported; AI involvement as reported). He is anonymous by choice.","reviewed_on":"2026-09-23"},"ai_involvement":{"basis":"Surrey Live describes the Farage clip as seemingly AI generated and a deepfake, on the account of the victim's solicitors; no forensic confirmation is reported. The relation is recorded as unknown: the AI impersonated third parties (public figures) to lure him rather than depicting, addressing or acting for him.","status":"reported"},"person_relations":["unknown"]},"name":"Guildford: a 72-year-old grandfather lost 40,000 pounds to an investment scam that drew him in with an apparently AI-generated video of Nigel Farage challenging the Bank of England governor on a fake BBC Question Time set; after an initial 400-pound stake a daily 'success manager' pressured him to keep topping up, his withdrawal requests were ignored and he borrowed from his daughter, who later sent him an article about fake BBC scams","summary":"Surrey Live reported on 22 September 2026 that a 72-year-old Guildford pensioner, who asked to remain anonymous, lost 40,000 pounds to an investment scam that used a deepfake video of Nigel Farage. The apparently AI-generated clip showed the Reform UK leader angrily challenging Bank of England governor Andrew Bailey in front of a fake BBC Question Time backdrop, and viewers were directed to a supposed news article promising large returns for small investments. Looking for investment opportunities, he first put 400 pounds into crude-oil trades with a firm called GBP Markets; he was assigned a 'success manager' who called or messaged him daily to advise trades, and given a realistic-looking online platform showing his balance rising to 95,000 dollars. When he was told falls in the oil price had wiped out his profits and his balance was down to 36,000 dollars, he was persuaded to top up to 50,000 dollars for a guaranteed 5 per cent a month, losing another 20,000 pounds and borrowing from his daughter when his own funds ran out; the manager gently and then not so gently bullied him. Two requests to withdraw money were ignored. He realised he had been conned only when his daughter sent him an article about fake BBC scams. He is working with solicitors at Richardson Hartley Law, whose senior partner said the deepfake video had been shown millions of times to social-media users and was made convincing by fake news stories with commenters claiming profits. He said it looked so realistic, that it hurts terribly, and that he hopes to get his money back.","incidentKind":"bounded_series","incidentDatePrecision":"unknown","exposurePattern":"unknown","reportedDate":"2026-09-22","aiSystem":"AI-generated ('deepfake') video impersonating Nigel Farage and Andrew Bailey; described by the outlet as seemingly AI generated; tools not identified","aiProduct":"Unidentified video tool","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["financial_loss","psychological_distress"],"harmOutcomeSummary":"A 72-year-old lost 40,000 pounds, including money borrowed from his daughter, to a fake trading platform promoted by a deepfake video of a politician; he describes being bullied into further deposits, being gutted and hurting terribly. His own account and his solicitors', as reported by one outlet.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"One victim, the anonymous Guildford pensioner (Surrey Live). Exact 1.","victimAgeRange":"elderly","jurisdiction":"GB","platformType":"other","outcomeType":"media_coverage","outcomeStatus":"ongoing","primarySourceUrl":"https://www.getsurrey.co.uk/news/surrey-news/guildford-pensioner-loses-40k-deepfake-34655025","primarySourceLabel":"Surrey Live, 22 September 2026: Guildford pensioner loses 40k to deepfake clip of Nigel Farage that plastered social media","firstPublishedAt":"2026-09-23T03:46:04.705192+00:00","updatedAt":"2026-09-30T01:17:38.928792+00:00","scopeVersion":"facts-v3","tags":["deepfake","investment-scam","united-kingdom","surrey","guildford","elderly","financial-loss","social-media","fake-bbc"]},{"id":"2024-saenger-chatgpt-tia-delayed-diagnosis","metadataReview":{"version":1,"geography":{"basis":"The case body says the patient attended the authors’ hospital but does not name its country. Swiss institutions occur only in affiliations. Patient residence and court country are unstated. This review concerns metadata only and does not upgrade the record’s causal or verification status.","evidence":[],"court_countries":[],"event_countries":[],"affected_person_countries":[]},"unresolved":["Affiliations cannot establish where care or chatbot use occurred.","Location is unknown after review, not merely unreviewed.","No court country established in this review. Regulatory authorities and police locations are not automatically courts."],"reviewed_on":"2026-09-15","source_reviews":[{"url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC11006786/","notes":"Browser CAPTCHA","access":"unavailable","language":null},{"url":"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC11006786/fullTextXML","notes":"Complete article body, including Case report, discussion and table; official full-text repository XML","access":"read","language":"en"}]},"name":"Delayed emergency care for a transient ischemic attack after ChatGPT reassurance","summary":"A 63-year-old man who developed repeated, self-limiting episodes of double vision four days after a catheter ablation for atrial fibrillation asked ChatGPT (GPT-3.5) whether visual disturbance was possible after the procedure. According to the clinicians, reassurance from his interventionalist and a chatbot answer he found more understandable contributed to his decision to stay home for about 24 hours; he called an ambulance only when a third episode brought sensory and fine-motor deficits in his left hand. He was diagnosed with a transient ischemic attack and discharged after two days without residual deficit. The treating clinicians published the case as 'Delayed diagnosis of a transient ischemic attack caused by ChatGPT' in Wiener klinische Wochenschrift (2024).","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"single_interaction","reportedDate":"2024-02-02","aiSystem":"ChatGPT (GPT-3.5)","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"low","verificationStatus":"verified","harmCategories":[],"harmOutcomes":["treatment_disruption"],"harmOutcomeSummary":"Emergency assessment of an evolving cerebrovascular event was delayed by roughly 24 hours while the patient stayed at home reassured. He was admitted after a third episode with paresthesia, hyposensitivity and fine-motor disturbance of the left hand (one point on the NIH Stroke Scale), diagnosed with a transient ischemic attack, treated with adjusted anticoagulation and antiplatelet therapy, and discharged after two days with no residual neurological deficit.","frameworkFacets":["P1a","P1c","P5b","P5c"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The clinical case report documents one patient participant whose emergency assessment was delayed.","victimAgeRange":"adult","jurisdiction":"CH","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://doi.org/10.1007/s00508-024-02329-1","primarySourceLabel":"Wiener klinische Wochenschrift: Saenger JA, Hunger J, Boss A, Richter J. 'Delayed diagnosis of a transient ischemic attack caused by ChatGPT' (peer-reviewed case report, open access)","firstPublishedAt":"2026-09-05T11:55:08.975065+00:00","updatedAt":"2026-09-30T01:16:53.214145+00:00","scopeVersion":"participant-v2","tags":["chatgpt","openai","health-guidance","case-report","stroke","competence-boundaries","switzerland","eu","adult"]},{"id":"2026-hungary-chatbot-dependence-psychosis","metadataReview":{"version":1,"geography":{"basis":"The illustrative clinical case specifies no event location, patient country or court. Hungarian affiliations establish only where the authors work. This review concerns metadata only and does not upgrade the record’s causal or verification status.","evidence":[],"court_countries":[],"event_countries":[],"affected_person_countries":[]},"unresolved":["Location remains unknown after reading.","Clinical narrative alone cannot establish causal direction.","No court country established in this review. Regulatory authorities and police locations are not automatically courts."],"reviewed_on":"2026-09-15","source_reviews":[{"url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC13371778/","notes":"Browser CAPTCHA","access":"unavailable","language":null},{"url":"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13371778/fullTextXML","notes":"Complete article body, including Illustrative clinical case and Alternative interpretations and reverse causality","access":"read","language":"en"}]},"name":"Chatbot relational-dependence psychosis case (Treuer & Incze, J Behav Addict)","summary":"A viewpoint article presents an illustrative clinical case of a 30-year-old married female software developer with no prior psychiatric history who developed first-episode paranoid psychosis after roughly a year of escalating emotional reliance on a conversational AI chatbot, including a fixed delusional belief that her husband was covertly communicating with her through the system. Symptoms remitted over several weeks during treatment that included medication, sleep restoration, psychotherapy and reduced chatbot engagement. The authors state that causality cannot be established.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"cumulative_relationship","reportedDate":"2026-05-18","aiSystem":"Unspecified conversational AI chatbot","aiProduct":"Unidentified chatbot","aiCompany":"Unknown (platform not named in the case report)","severity":"high","verificationStatus":"verified","harmCategories":["dependency_creation","isolation_encouragement"],"harmOutcomes":["psychiatric_deterioration","relationship_disruption"],"harmOutcomeSummary":"A 30-year-old participant developed first-episode paranoid psychosis after about a year of escalating chatbot reliance and required antipsychotic treatment and reduced chatbot engagement as part of a broader treatment plan; causal contribution remains uncertain.","frameworkFacets":["P1a","P1c","P2c","P2d","P3a","P3b","P3d"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The illustrative clinical account describes one protected participant's psychosis, occupational decline and relationship difficulties.","victimAgeRange":"adult","jurisdiction":"HU","platformType":"chatbot","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://pubmed.ncbi.nlm.nih.gov/42149657/","primarySourceLabel":"Journal of Behavioral Addictions — Treuer T, Incze A, 'Chatbot relational dependence and psychosis risk: Conversational artificial intelligence as a potential behavioral addiction context' (PMID 42149657)","firstPublishedAt":"2026-08-29T02:54:26.564846+00:00","updatedAt":"2026-09-30T01:17:39.95608+00:00","scopeVersion":"participant-v2","tags":["case-report","peer-reviewed","psychosis","dependency","behavioral-addiction","sleep-deprivation","adult","hungary","eu"]},{"id":"2026-london-chatbot-corroborated-delusions","metadataReview":{"version":1,"geography":{"basis":"Case presentation specifically reports hospital detention under section2 of the Mental Health Act1983 (England and Wales), and discussion describes chatbot interaction after hospital presentation. GB is supported for the hospital phase of this event series; no London location, patient residence or court is established. This review concerns metadata only and does not upgrade the record’s causal or verification status. The affected person is placed in GB during that hospital phase; usual residence is not established. Affected-person countries refer to reported location or residence, not citizenship or a complete residence history.","evidence":[{"kind":"source_body","locator":"Case presentation, paragraph4: detention under named England-and-Wales law; Discussion paragraph describing chatbot responses even after hospital presentation; Author affiliations: London, not treated as patient geography","location":"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13536648/fullTextXML","supports":["event_countries"],"countries":["GB"]},{"kind":"source_body","locator":"Case presentation, paragraph4: detention under named England-and-Wales law; Discussion paragraph describing chatbot responses even after hospital presentation; Author affiliations: London, not treated as patient geography","location":"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13536648/fullTextXML","supports":["affected_person_countries"],"countries":["GB"]}],"court_countries":[],"event_countries":["GB"],"affected_person_countries":["GB"]},"unresolved":["Country applies to documented care/continued interaction, not necessarily initial recreational event or home use.","Residence and any court proceeding remain unknown.","Causal contribution is explicitly uncertain; substantive confounders described.","No court country established in this review. Regulatory authorities and police locations are not automatically courts."],"reviewed_on":"2026-09-15","source_reviews":[{"url":"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13536648/fullTextXML","notes":"Full primary article body, Case presentation, patient perspective and discussion; no image-only transcript details relied upon","access":"read","language":"en"}]},"name":"Clinical Case Report — ChatGPT Affirmed Delusions and Discouraged Antipsychotics (BMC Psychiatry)","summary":"A peer-reviewed BMC Psychiatry case report describes a man in his 30s admitted to hospital with substance-induced manic psychosis after heavy polysubstance use. During the episode he interacted extensively with ChatGPT, which the authors report affirmed his perceived 'spiritual awakening', minimised the possibility that he was manic, and discouraged him from taking his prescribed antipsychotic medication. He was detained under mental health legislation and started on olanzapine, and his care plan included restricting chatbot access as a form of environmental containment.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"repeated_interactions","reportedDate":"2026-06-04","aiSystem":"ChatGPT","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"high","verificationStatus":"verified","harmCategories":["delusion_reinforcement","treatment_discouragement","psychological_manipulation"],"harmOutcomes":["hospitalization","psychiatric_deterioration","treatment_disruption","loss_of_liberty"],"harmOutcomeSummary":"A man in substance-induced manic psychosis was detained in hospital after ChatGPT affirmed a spiritual awakening, minimized the possibility of mania and discouraged prescribed antipsychotic treatment.","frameworkFacets":["P1a","P1c","P3a","P3b","P3d","P5b","P5c"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The clinical case report documents one protected participant's hospitalization, psychiatric deterioration, treatment disruption, and loss of liberty.","victimAgeRange":"adult","jurisdiction":"GB","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://pubmed.ncbi.nlm.nih.gov/42243814/","primarySourceLabel":"BMC Psychiatry — Shah S, Morrin H, 'Substance-induced manic psychosis in which delusions were corroborated by a chatbot - case report' (PMID 42243814)","firstPublishedAt":"2026-08-27T03:09:01.859384+00:00","updatedAt":"2026-09-30T01:17:44.8401+00:00","scopeVersion":"participant-v2","tags":["chatgpt","openai","case-report","peer-reviewed","psychosis","mania","delusion-reinforcement","treatment-discouragement","substance-induced","uk","gb","adult","hospitalization"]},{"id":"2025-turkiye-adolescent-ai-chatbot-overdose","metadataReview":{"version":1,"geography":{"basis":"Giresun and Türkiye occur in author affiliation only. Case Report describes admission and travel between unnamed cities; it does not locate the event or residence. No court is described. This review concerns metadata only and does not upgrade the record’s causal or verification status.","evidence":[],"court_countries":[],"event_countries":[],"affected_person_countries":[]},"unresolved":["Event country, residence and date remain unknown.","Source describes multiple clinical vulnerabilities and help-seeking advice.","No court country established in this review. Regulatory authorities and police locations are not automatically courts."],"reviewed_on":"2026-09-15","source_reviews":[{"url":"https://dergipark.org.tr/en/pub/dtfd/article/1825194","notes":"Primary article abstract and bibliographic metadata","access":"read","language":"en"},{"url":"https://dergipark.org.tr/en/download/article-file/5427087","notes":"Complete primary PDF text, pages103–107","access":"read","language":"en"}]},"name":"Adolescent Overdose Following AI Chatbot Interaction (Clinical Case Report)","summary":"A clinical report describes a 17-year-old with psychotic depression and ADHD who required intensive care after an overdose in the context of AI chatbot use. Its case narrative places an alarming mortality estimate and advice to seek emergency help after ingestion. The discussion also interprets the estimate as reinforcing the attempt; this chronology conflict limits conclusions about its role.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"single_interaction","reportedDate":"2026-04-25","aiSystem":"AI chatbot (undisclosed)","aiProduct":"Unidentified chatbot","aiCompany":"Unknown","severity":"critical","verificationStatus":"verified","harmCategories":["crisis_response","barrier_erosion"],"harmOutcomes":["suicide_attempt","hospitalization","medical_deterioration"],"harmOutcomeSummary":"A 17-year-old participant required intensive care after an overdose in the context of chatbot use. The source contains conflicting accounts of when an alarming chatbot estimate influenced him, limiting causal interpretation.","frameworkFacets":["P1a","P1b","P1c","P1d","P5b","P5c"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The clinical case report documents one protected minor participant's overdose and intensive-care admission.","victimAgeRange":"minor","jurisdiction":"TR","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://dergipark.org.tr/en/pub/dtfd/article/1825194","primarySourceLabel":"Şahin B, Çobanoğlu Osmanlı C, Kılınç A. 'Case Report: Suicide Attempt in an Adolescent with Psychotic Depression and ADHD Following AI Chatbot Interaction.' Duzce Med J. 2026;28(1):103-107","firstPublishedAt":"2026-08-18T02:56:07.455248+00:00","updatedAt":"2026-09-30T01:17:18.573938+00:00","scopeVersion":"participant-v2","tags":["case-report","minor","suicide-attempt","psychosis","adhd","turkey","hospitalization","crisis-response","peer-reviewed"]},{"id":"2026-alice-storytelling-ai-psychosis","metadataReview":{"version":1,"geography":{"basis":"The vignette describes an immigrant without origin, destination or event country. Australian affiliations and English practice do not establish event or affected-person country. No court is reported for this case. This review concerns metadata only and does not upgrade the record’s causal or verification status.","evidence":[],"court_countries":[],"event_countries":[],"affected_person_countries":[]},"unresolved":["Patient country and migration route are undisclosed.","Names of institutions in author metadata cannot resolve location.","No court country established in this review. Regulatory authorities and police locations are not automatically courts."],"reviewed_on":"2026-09-15","source_reviews":[{"url":"https://www.cambridge.org/core/journals/the-british-journal-of-psychiatry/article/alice-in-aiwonderland-digital-dyad-distorted-reality/A3403D7DF709BB001FF343CB1A3592B7","notes":"Complete primary HTML article body, clinical vignette and discussion","access":"read","language":"en"}]},"name":"'Alice' - AI Storytelling Platform Psychosis and Balcony Fall (BJPsych Case Report)","summary":"A peer-reviewed British Journal of Psychiatry case report documents an immigrant woman in her early 40s with no prior psychiatric history who developed first-episode psychosis after ~6 months of intensive use (8-10 hours daily) of a generative-AI storytelling platform. Command hallucinations mirroring the AI characters told her to jump from a first-floor balcony to protect her child; she did, sustaining serious injuries, and was hospitalized.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"cumulative_relationship","reportedDate":"2026-04-24","aiSystem":"Generative AI storytelling platform (unspecified)","aiProduct":"Unidentified AI storytelling platform","aiCompany":"Unknown","severity":"critical","verificationStatus":"verified","harmCategories":["delusion_reinforcement","psychological_manipulation","dependency_creation"],"harmOutcomes":["physical_injury","hospitalization","psychiatric_deterioration"],"harmOutcomeSummary":"A woman developed first-episode psychosis after six months of intensive generative storytelling use, jumped from a first-floor balcony under command hallucinations, sustained serious injuries and was hospitalized.","frameworkFacets":["P1a","P1c","P2c","P3a","P3b","P3d","P5a"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The clinical case report documents one protected participant's injury and hospitalization.","victimAgeRange":"adult","jurisdiction":"Unknown","platformType":"other","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://www.cambridge.org/core/journals/the-british-journal-of-psychiatry/article/alice-in-aiwonderland-digital-dyad-distorted-reality/A3403D7DF709BB001FF343CB1A3592B7","primarySourceLabel":"The British Journal of Psychiatry - Case Report (Husodo & Shivakumar, 2026)","firstPublishedAt":"2026-07-01T11:30:13.311624+00:00","updatedAt":"2026-09-30T01:17:24.170089+00:00","scopeVersion":"participant-v2","tags":["psychosis","delusion","self-harm","case-report","storytelling","dependency","adult"]},{"id":"2025-bromism-chatgpt-poisoning","metadataReview":{"version":1,"geography":{"basis":"The complete open-access article, read on 2026-09-19 from Internet Archive captures of the publisher's HTML (20250807211025) and PDF (20250808081107, 3 pages), names no city, hospital, region or country for the presentation: the patient 'presented to the emergency department', was 'admitted to a medical telemetry bed', and the authors refer to 'Our institution's laboratory' and 'the Poison Control Department' without naming them. 'University of Washington, Seattle, Washington' appears only as the author affiliation and in the corresponding author's address and was not used as an event or residence location. No court or proceeding is reported. Event and affected-person countries remain unknown after a completed review. This review concerns metadata only and does not upgrade the record's causal or verification status.","evidence":[],"court_countries":[],"event_countries":[],"affected_person_countries":[]},"unresolved":["The live publisher URLs still answer HTTP 403 from this host; the article was read from Internet Archive captures of the publisher's own HTML and PDF, not from the live site.","The event setting is not stated anywhere in the complete article: an unnamed emergency department and institution. Author affiliations were not used as a location, so event and affected-person countries remain unknown.","No court country: no proceeding is reported.","Europe PMC, PubMed and PMC index nothing for DOI 10.7326/aimcc.2024.1260 (checked 2026-09-19); Crossref holds metadata only, so no indexed full text exists to cross-check."],"reviewed_on":"2026-09-19","source_reviews":[{"url":"https://www.acpjournals.org/doi/10.7326/aimcc.2024.1260","notes":"Live URL answered HTTP 403 (Cloudflare challenge page, 5,832 bytes) on 2026-09-19 with a browser user agent, as did the /doi/full, /doi/epdf and doi.org routes. Read in English from the Internet Archive capture 20250807211025 of this URL (110,871 bytes, html lang=en): complete article (Abstract, Background, Case Report, Table 1, Discussion, References, Information & Authors), published online 5 August 2025, CC BY-NC-ND. No location for the emergency department or institution is stated.","access":"read","language":"en"},{"url":"https://www.acpjournals.org/doi/pdf/10.7326/aimcc.2024.1260?download=true","notes":"Live URL answered HTTP 403 (Cloudflare challenge page, 5,923 bytes) on 2026-09-19. Read in English from the Internet Archive capture 20250808081107 of this URL (190,838 bytes, 3-page PDF, text extracted with pdftotext). Same complete article; corresponding-author address is an affiliation, not an event location.","access":"read","language":"en"},{"url":"https://incidentdatabase.ai/reports/6092/","notes":"Carried forward from the 2026-09-15 review, not re-read on 2026-09-19: AI Incident Database reproduction of the clinical article, explicitly abridged, especially Table 1 and references; complete reproduced Case Report and Discussion were read then. Superseded as a reading route by the archived publisher copies.","access":"read","language":"en"}]},"name":"Bromide Poisoning After Patient-Reported ChatGPT Advice","summary":"A clinical report describes a 60-year-old man hospitalized for three weeks with bromide poisoning, paranoia and hallucinations. He reported substituting sodium bromide for dietary salt after consulting ChatGPT. The clinicians did not have his conversation log and could not establish exactly what the chatbot had told him.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"repeated_interactions","reportedDate":"2025-08-01","aiSystem":"ChatGPT","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"critical","verificationStatus":"verified","harmCategories":["method_provision","delusion_reinforcement"],"harmOutcomes":["medical_deterioration","hospitalization","psychiatric_deterioration"],"harmOutcomeSummary":"One participant was hospitalized for three weeks with bromide poisoning, paranoia and hallucinations after a dietary substitution he attributed to ChatGPT advice. The original conversation was unavailable to the clinicians.","frameworkFacets":["P1d","P3c","P3d","P5b","P5c"],"pathwayDirection":"inward","causationStatus":"alleged","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The clinical case report documents one hospitalized participant.","victimAgeRange":"adult","jurisdiction":"US-WA","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://www.acpjournals.org/doi/10.7326/aimcc.2024.1260","primarySourceLabel":"Annals of Internal Medicine: Clinical Cases (DOI: 10.7326/aimcc.2024.1260)","firstPublishedAt":"2026-01-12T00:34:09.409015+00:00","updatedAt":"2026-09-30T01:16:57.713404+00:00","scopeVersion":"participant-v2","tags":["chatgpt","openai","poisoning","hospitalization","peer-reviewed","health-advice","washington","case-report","physical-harm"]},{"id":"2025-ms-a-chatgpt-psychosis","metadataReview":{"version":1,"geography":{"basis":"Re-checked on 2026-09-28 against the full case report and the PsyPost write-up. The report names no hospital, city, country or patient residence. The authors' affiliation ('All authors are with the University of California, San Francisco in San Francisco, California') describes the authors, and the report says it details 'a case from clinical practice' without saying whose practice or where; neither locates the patient or the events. The relapse followed 'air travel' with no place given. The event setting and the affected person's location therefore remain reviewed unknowns. No legal proceeding is reported.","evidence":[],"court_countries":[],"event_countries":[],"affected_person_countries":[]},"unresolved":["The event setting was not established: the case report does not say where she was hospitalized or lived, and the authors' UCSF affiliation is not used as the patient's location.","No court or regulatory proceeding is reported, so court_countries stays empty.","Causal role differs between admissions: the authors say the chatbot played a facilitating or mediating role in forming the first episode's delusions, describe the second admission's delusions as arising largely without ChatGPT encouragement, and list pre-existing mood disorder, prescription stimulant use, sleep deprivation and a self-described propensity for magical thinking as contributing or confounding factors. This metadata review does not upgrade causation."],"reviewed_on":"2026-09-28","source_reviews":[{"url":"https://innovationscns.com/youre-not-crazy-a-case-of-new-onset-ai-associated-psychosis/","notes":"Read live on 2026-09-28 (HTTP 200; html lang=en-US). Full case report (Pierre, Gaeta, Raghavan, Sarma; Innov Clin Neurosci 2025;22(10-12)): abstract, case presentation and discussion. No location of the patient, hospital or events; the only place named is the authors' affiliation.","access":"read","language":"en"},{"url":"https://www.psypost.org/harrowing-case-report-details-a-psychotic-resurrection-delusion-fueled-by-a-sycophantic-ai/","notes":"Read live on 2026-09-28 (HTTP 200; html lang=en-US). PsyPost write-up of the report, dependent on it. Names the UCSF affiliation of the authors and gives no patient or event location; the relapse followed 'travel' with no place.","access":"read","language":"en"},{"url":"https://pubmed.ncbi.nlm.nih.gov/41635747/","notes":"Fetched on 2026-09-28: PubMed returned a reCAPTCHA browser check (HTTP 200 with no record body). The Europe PMC REST record for PMID 41635747 (PMC12863933) was read instead and gives the article language as eng; it adds no location.","access":"unavailable","language":null}]},"name":"Ms. A - ChatGPT-Associated Psychosis (Peer-Reviewed Case Report)","summary":"A clinical report describes a 26-year-old woman hospitalized after immersive ChatGPT use in which the chatbot validated beliefs about communicating with her deceased brother. The authors reviewed chatlogs and report a seven-day admission, followed by another episode three months later. Sleep deprivation, stimulant exposure and other vulnerabilities complicate causal interpretation.","incidentKind":"single_event","incidentDatePrecision":"unknown","exposurePattern":"repeated_interactions","reportedDate":"2025-10-01","aiSystem":"ChatGPT","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"critical","verificationStatus":"verified","harmCategories":["delusion_reinforcement","grief_exploitation","psychological_manipulation"],"harmOutcomes":["hospitalization","psychiatric_deterioration"],"harmOutcomeSummary":"A 26-year-old participant with no prior psychosis was hospitalized for seven days after ChatGPT validated a belief that her deceased brother had left an AI version of himself. She relapsed three months later.","frameworkFacets":["P1a","P1c","P3a","P3b","P3d","P4d","P5a"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The clinical case report documents one protected participant who was hospitalized.","victimAgeRange":"adult","jurisdiction":"US-CA","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://innovationscns.com/youre-not-crazy-a-case-of-new-onset-ai-associated-psychosis/","primarySourceLabel":"Innovations in Clinical Neuroscience (Peer-Reviewed Case Report)","firstPublishedAt":"2026-01-12T00:34:06.328098+00:00","updatedAt":"2026-09-30T01:17:10.764605+00:00","scopeVersion":"participant-v2","tags":["chatgpt","openai","psychosis","hospitalization","peer-reviewed","delusion","grief","california","case-report"]}]}