{"meta":{"exportedAt":"2026-09-30T06:29:35.124Z","formatVersion":2,"selection":{"q":"evidence","system":"","harm":"","context":"","country":"","role":"","relation":"","evidence":"","year":"2026","response":"","severity":"","verification":"","view":"incidents","sort":"added"},"totalIncidents":6,"coverage":{"cases":6,"countries":4,"languages":2,"unknownLocation":1,"locationPending":0,"unknownLanguage":1,"unknownDate":0,"lawsuits":1,"regulatory":0,"minors":1,"coreRelations":3,"contextualRelations":1,"mixedRelations":0,"unknownRelations":2,"relationPending":1,"relationUnknown":1},"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":"2026-sdny-heppner-claude-defense-strategy-documents-ruled-not-privileged","caseFacts":{"claims":[{"id":"c1","status":"documented","evidence":[{"locator":"[Page 1] 'At a pretrial conference in this matter held on February 10, 2026, the Court, after hearing the arguments of counsel, granted from the bench the Government's motion for a ruling that certain written exchanges that defendant Bradley Heppner had with a generative artificial intelligence (\"AI\") platform were not protected from Government inspection by either the attorney-client privilege or the work product doctrine.'","relation":"supports","source_id":"s1"},{"locator":"[Page 12] 'Because Heppner's use of Claude fails to satisfy either of these rules, the AI Documents do not merit the protections Heppner has claimed.'","relation":"supports","source_id":"s1"},{"locator":"'But the judge said that protections for conversations between clients and their attorneys did not apply: Claude is not a lawyer, and the executive's attorneys did not ask him to talk to the chatbot.'","relation":"supports","source_id":"s2"}],"assertion":"On 10 February 2026 the court granted from the bench the Government's motion for a ruling that Heppner's written exchanges with a generative AI platform were not protected from Government inspection by either attorney-client privilege or the work product doctrine; a memorandum of reasons followed on 17 February 2026.","causal_attribution":"Established by the court's own memorandum."},{"id":"c2","status":"documented","evidence":[{"locator":"[Page 3] 'In connection with Heppner's arrest on November 4, 2025, agents with the Federal Bureau of Investigation executed a search warrant at Heppner's home and seized numerous documents and electronic devices.'; 'Heppner's counsel later represented to the Government that among the seized materials were approximately thirty-one documents that memorialize communications that Heppner had with the generative AI platform \"Claude,\" which is operated by the private company Anthropic.'","relation":"supports","source_id":"s1"},{"locator":"'A criminal suspect consulted the Claude artificial intelligence tool, opening 31 separate chats and entering sensitive details about his case, including information he had received from his attorneys. The FBI later raided his home and seized printouts of the conversations.'","relation":"context","source_id":"s3"}],"assertion":"FBI agents seized about thirty-one documents memorialising Heppner's communications with Claude when they executed a search warrant at his home in connection with his arrest in November 2025 (the memorandum gives 4 November on page 3 and the day after the 4 November unsealing on page 2).","causal_attribution":"Court record of counsel's representation; not a claim about AI behaviour."},{"id":"c3","status":"documented","evidence":[{"locator":"[Page 3] 'Without any suggestion from counsel that he do so, Heppner \"prepared reports that outlined defense strategy, that outlined what he might argue with respect to the facts and the law that we anticipated that the government might be charging.\"'; [Page 4] 'Heppner's counsel conceded, however, that counsel \"did not direct [Heppner] to run Claude searches.\"'","relation":"supports","source_id":"s1"}],"assertion":"According to his counsel, Heppner used Claude in 2025, after receiving a grand jury subpoena, to prepare reports outlining defence strategy, without any suggestion from counsel.","causal_attribution":"The memorandum records counsel's own statements."},{"id":"c4","status":"documented","evidence":[{"locator":"[Page 5] 'Because Claude is not an attorney, see ECF No. 23-6, that alone disposes of Heppner's claim of privilege.'; [Page 6] 'Second, the communications memorialized in the AI Documents were not confidential.'","relation":"supports","source_id":"s1"}],"assertion":"The court held that the documents were not privileged because Claude is not an attorney and the exchanges were not confidential.","causal_attribution":"Established by the court's reasoning."},{"id":"c5","status":"reported","evidence":[{"locator":"'In May, Heppner was convicted of securities fraud and wire fraud, among other charges.'","relation":"supports","source_id":"s2"}],"assertion":"Heppner was convicted in May 2026 of securities fraud and wire fraud, among other charges.","causal_attribution":"No inspected source attributes the conviction to the Claude documents or says they were used at trial."}],"effects":[{"label":"the defendant's self-prepared defence-strategy documents, written with Claude, were ruled open to Government inspection","claim_id":"c1","direction":"negative"}],"sources":[{"id":"s1","url":"https://storage.courtlistener.com/recap/gov.uscourts.nysd.652137/gov.uscourts.nysd.652137.27.0.pdf","kind":"court_record","access":"read","language":"en","translation_note":"Scanned PDF without a text layer; the 12 page images were read on 2026-09-29 and the passages relied on were transcribed manually (bodies/law-heppner-memo27-transcribed.txt).","independence_group":"sdny-memo"},{"id":"s2","url":"https://www.yahoo.com/news/us/articles/confided-chatgpt-secrets-ended-court-090000568.html","kind":"news_report","access":"read","language":"en","translation_note":"Washington Post, 27 August 2026, read in full via its Yahoo News syndication on 2026-09-29. Its account of the ruling derives from the court record; the conviction report is its own.","independence_group":"washington-post"},{"id":"s3","url":"https://www.i24news.tv/en/news/international/technology-science/artc-courtroom-hallucinations-lawyers-cite-precedents-dreamt-up-by-ai-as-clients-sabotage-their-own-cases","kind":"news_report","access":"read","language":"en","translation_note":"Read via a Wayback capture (2026-09-10) on 2026-09-29; the live page is a JavaScript shell. The article does not name the defendant; the identification rests on the matching court, platform and ruling.","independence_group":"i24news"},{"id":"s4","url":"https://www.abc.net.au/news/2026-09-10/christopher-duff-to-stand-trial-over-nsw-ai-court-data-breach/107135032","kind":"news_report","access":"read","language":"en","translation_note":"ABC News (Australia), 10 September 2026, read in full; mentions the New York ruling only as a precedent cited by a police prosecutor.","independence_group":"abc-au"}],"version":1,"ai_roles":["own_use"],"contexts":["justice"],"unknowns":["Whether the Government used the Claude documents at trial or in plea or sentencing proceedings.","What the documents contained beyond counsel's description as defence-strategy reports.","Whether any appeal challenges the privilege ruling."],"geography":{"basis":"The ruling was made by the U.S. District Court for the Southern District of New York, and the memorandum places his arrest in the Northern District of Texas. Where he used Claude and his country of residence are not stated.","court_countries":["US"],"event_countries":["US"],"affected_person_countries":[]},"publication":{"basis":"Published as a documented legal consequence of a person's AI use: a federal court ruled that a defendant's self-prepared defence documents written with Claude were open to Government inspection. The record rests on the court's memorandum; the conviction is recorded as context and is not attributed to the AI use.","reviewed_on":"2026-09-29"},"ai_involvement":{"basis":"The court memorandum records counsel's representation that the seized documents memorialise Heppner's communications with Claude and that he used it to prepare defence-strategy reports.","status":"supported"},"person_relations":["communicated_with"]},"name":"S.D.N.Y.: Judge Rakoff rules that a fraud defendant's written exchanges with Claude about his defence, seized by the FBI, are protected by neither attorney-client privilege nor work product","summary":"Bradley Heppner, a corporate executive charged in the Southern District of New York with securities fraud, wire fraud and related offences, used Anthropic's Claude in 2025, after receiving a grand jury subpoena, to prepare reports outlining his defence strategy. His counsel had not directed him to do so. FBI agents seized about thirty-one documents memorialising those exchanges when they searched his home in connection with his arrest in November 2025. Heppner claimed privilege over them. On 10 February 2026 Judge Jed S. Rakoff granted from the bench the Government's motion for a ruling that the documents were not protected from Government inspection by either the attorney-client privilege or the work product doctrine, and a memorandum filed 17 February 2026 gave the reasons: Claude is not an attorney, and the communications were not confidential under Anthropic's privacy policy. The court described the question as one of first impression nationwide. Heppner was convicted in May 2026 of securities fraud, wire fraud and other charges; no inspected source says whether the Claude documents were used at trial.","incidentDate":"2026-02-10","incidentKind":"single_event","incidentDatePrecision":"day","exposurePattern":"repeated_interactions","reportedDate":"2026-02-17","aiSystem":"Claude (Anthropic), a publicly available generative AI platform the defendant used on his own to prepare defence-strategy reports","aiProduct":"Claude","aiCompany":"Anthropic","severity":"low","verificationStatus":"verified","harmCategories":[],"harmOutcomes":["legal_harm"],"harmOutcomeSummary":"The defendant's self-prepared defence-strategy documents, written with Claude, lost the protection he claimed and were ruled open to Government inspection in his criminal case (court memorandum). Whether they were used at his trial is not reported.","frameworkFacets":[],"causationStatus":"established","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"One person: Heppner, the Claude user whose documents were ruled unprotected (court memorandum). No other person's harm from the AI use is reported; investors' fraud losses are the subject of the charges and are not counted here.","victimAgeRange":"adult","jurisdiction":"US-NY","platformType":"assistant","outcomeType":"criminal_charges","outcomeStatus":"resolved","primarySourceUrl":"https://storage.courtlistener.com/recap/gov.uscourts.nysd.652137/gov.uscourts.nysd.652137.27.0.pdf","primarySourceLabel":"United States v. Heppner, S.D.N.Y. 25 Cr. 503 (JSR), Memorandum (ECF No. 27), filed 17 February 2026","firstPublishedAt":"2026-09-29T09:10:12.292608+00:00","updatedAt":"2026-09-30T01:17:52.181801+00:00","scopeVersion":"facts-v3","tags":["justice","privilege","evidence","claude","anthropic","own-use","criminal-case","sdny","first-impression"]},{"id":"2026-jeju-middle-school-student-filmed-teachers-deepfake-composites-referred","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'제주동부경찰서는 도내 모 중학교 재학생 A군을 성폭력범죄의 처벌 등에 관한 특례법 위반(카메라 등 이용 촬영) 혐의로 지난 17일 불구속 송치했다고 23일 밝혔다.'","relation":"supports","source_id":"s1"},{"locator":"'제주동부경찰서는 성폭력범죄의 처벌 등에 관한 특례법상 카메라 등을 이용한 촬영 혐의로 A 군을 불구속 송치했다고 23일 밝혔다.'","relation":"supports","source_id":"s2"}],"assertion":"Jeju Dongbu Police Station announced on 23 September 2026 that a student at a Jeju middle school had been referred to prosecutors without detention on 17 September for violating the Sexual Violence Punishment Act (filming using a camera or similar device).","causal_attribution":"Police announcement carried by both outlets (one origin)."},{"id":"c2","status":"reported","evidence":[{"locator":"'경찰은 A군의 휴대전화를 보호자 측으로부터 임의제출받아 디지털 포렌식을 진행했다. 포렌식 과정에서 일부 교사와 관련된 촬영물과 일부 성적 합성물이 확인됐으며, A군은 범행을 인정하고 자백한 것으로 파악됐다.'; '포렌식 결과 일부 선정적인 합성물이 확인됐고, 수치심을 유발할 정도의 자료는 범죄 혐의에 포함해 수사했다'","relation":"supports","source_id":"s1"}],"assertion":"The student's phone was voluntarily submitted by his guardians and forensically examined; the examination found footage relating to some teachers and some sexual composites, the student admitted the conduct, and police say suggestive composites causing shame were included in the charged conduct.","causal_attribution":"Police account to Seoul Shinmun."},{"id":"c3","status":"reported","evidence":[{"locator":"'처음 신고된 피해 교사 외에도 피해 교사가 더 있는 것은 맞지만 몇 명인지는 밝히기 어렵다'; '일부에서 학생 피해자가 있다는 이야기가 있었지만 확인 결과 해당 학교 재학생 피해자는 없었다'; '피해교사들이 현재도 해당 학교에서 학생들을 가르치고 있어 피해 사실이 알려질 경우 2차 피해로 이어질 수 있다는 이유에서다.'","relation":"supports","source_id":"s1"},{"locator":"'피해자가 다수 있는 것을 확인했다'","relation":"supports","source_id":"s2"}],"assertion":"Police say there are more victim teachers than the one first reported but withheld the number because the teachers still teach at the school and disclosure could cause secondary harm, and that no student at the school was a victim.","causal_attribution":"Police official quoted by Seoul Shinmun."},{"id":"c4","status":"reported","evidence":[{"locator":"'특히 피해교사들이 직접 여러 학생과 관련된 촬영물을 확인했는데도 별도의 강제 압수수색 없이 사건이 마무리되는 것 아니냐는 우려를 제기했다.'; '임의제출받은 휴대전화에 대한 포렌식으로 필요한 증거를 확보했다'","relation":"supports","source_id":"s1"}],"assertion":"The Jeju teachers' union questioned whether the full scope of harm was established, saying the victim teachers had themselves seen footage involving several students yet no compulsory seizure was made; police replied that forensics of the voluntarily submitted phone secured the necessary evidence.","causal_attribution":"Union and police positions as reported."},{"id":"c5","status":"reported","evidence":[{"locator":"'피해교사들은 사건이 드러난 지 두 달이 지나도록 병원 치료와 심리상담을 받으면서도 매일 학교에 출근해 수업하고 있다.'; '지난 14일에는 변호사 조력 없이 경찰 피해자 조사를 받았고, 17일에는 교육청과 학교안전공제회의 조력 없이 지역교권보호위원회에 출석했다. 형사고소장도 피해교사들이 직접 작성해 제출했다고 노조는 주장했다.'; '대체교사를 구하기 어려워 사용할 수 있는 특별휴가 16일 가운데 3~4일 정도만 사용하고 있다는 게 노조 측 설명이다.'; '노조는 학교안전공제회가 형사고소 변호사비를 피해교사 개인별이 아닌'","relation":"supports","source_id":"s1"}],"assertion":"According to the union, two months after the case came to light the victim teachers are receiving hospital treatment and counselling while going to school every day; on 14 September they gave police statements without a lawyer and on 17 September attended the regional teachers' rights protection committee without institutional support; they wrote the criminal complaint themselves; they have used only three or four of 16 days of special leave because substitute teachers cannot be found; and the school safety mutual-aid association proposes legal fees of 3.3 million won for all victim teachers combined, payable after the case ends.","causal_attribution":"The union's account as reported; not independently confirmed."},{"id":"c6","status":"reported","evidence":[{"locator":"'교육청에도 치료비나 휴식, 병가 등 피해교사에 대한 지원이 필요하다는 취지로 이야기했다'; '피해교사들은 지난 1일과 3일 두 차례 교육감 면담을 요청했지만 받아들여지지 않았고'; '노조는 오는 10월 2일까지 전국 교사·시민을 대상으로 가해자 엄벌 탄원과 피해교사 지원을 요구하는 서명운동을 벌인다.'","relation":"supports","source_id":"s1"}],"assertion":"Police said they had told the education office that the teachers need support such as treatment costs, rest and sick leave; the victim teachers requested meetings with the superintendent on 1 and 3 September without success and a third written request was sent to the superintendent's office on 21 September; the union is collecting signatures until 2 October 2026 for severe punishment and victim support.","causal_attribution":"Police and union statements as reported; the meeting requests are described in a passage introduced by reference to the union side."},{"id":"c7","status":"reported","evidence":[{"locator":"'A 군은 7월 중순쯤 모 중학교 현장체험학습 도중 교사의 신체를 휴대전화로 촬영한 혐의로 입건됐다.'; '경찰은 휴대전화 조사 과정에서 범행이 수개월간 이어졌으며 피해자가 다수 있는 것을 확인했다.'; '다만 경찰은 합성물을 유포한 정황은 발견하지 못했다고 전했다.'","relation":"supports","source_id":"s2"}],"assertion":"News1 reports that the student was booked for filming a teacher's body with his phone during a school field trip around mid-July 2026, that the phone examination showed the offending had continued for several months with multiple victims, and that police found no indication the composites were distributed.","causal_attribution":"Police account as reported by News1."}],"effects":[{"label":"teachers secretly filmed by a student, with some footage turned into sexual deepfake composites","claim_id":"c2","direction":"negative"},{"label":"victim teachers in hospital treatment and counselling for two months while teaching daily; questioned without a lawyer, attended the protection committee without support and wrote the complaint themselves","claim_id":"c5","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.seoul.co.kr/news/society/accident/2026/09/23/20260923500094","kind":"news_report","access":"read","language":"ko","translation_note":"Read live in Korean on 2026-09-23 (Seoul Shinmun, 강동삼, published 11:18 KST 23 September 2026): police announcement, the paper's own call to a police official, the Jeju teachers' union's statements. Model reading; no human bilingual review.","independence_group":"seoul-shinmun"},{"id":"s2","url":"https://www.news1.kr/local/jeju/6300227","kind":"news_report","access":"read","language":"ko","translation_note":"Read in Korean on 2026-09-23 through the r.jina.ai reader proxy (live page 403); News1 Jeju, 고동명, 23 September 2026: the police announcement of the referral, the mid-July field-trip filming for which he was booked, several months of offending with multiple victims, and no indication of distribution of the composites. Model reading; no human bilingual review.","independence_group":"news1"}],"version":1,"ai_roles":["others_use"],"contexts":["education","work","justice"],"unknowns":["The number of victim teachers (police confirm more than one).","The filming for which the student was booked took place around mid-July 2026 during a school field trip (News1); police say the offending continued for several months, so the start date is unknown; the record uses July 2026 with month precision.","The tool used and whether the composites are separately charged.","The student's age and grade.","Whether prosecutors indict and how the juvenile case is handled.","Korean bodies were read by the model without human bilingual review."],"geography":{"basis":"A middle school on Jeju, South Korea; investigated by Jeju Dongbu Police Station; the teachers work there (Seoul Shinmun, News1). Referred to prosecutors; no court proceeding yet.","court_countries":[],"event_countries":["KR"],"affected_person_countries":["KR"]},"publication":{"basis":"Published under the 2026-09-15 charter as a harmful-artifact case in a school with an ongoing adverse experience for the victim teachers, documented by a police announcement and the paper's own police and union interviews; the student admitted the conduct per police. Single origin for each claim, so all are recorded as reported. No teacher, student or school is named; the student is a minor.","reviewed_on":"2026-09-23"},"ai_involvement":{"basis":"Police say forensic examination of the student's phone found sexual composites made from footage of teachers; both outlets describe them as deepfake composites and News1 describes the suspected method as deepfake; the tool is not named. The primary charge is illegal filming, and police told News1 they found no indication the composites were distributed.","status":"reported"},"person_relations":["depicted_or_impersonated"]},"name":"Jeju: a middle-school student who secretly filmed teachers at his school and turned some of the footage into sexual deepfake composites was referred to prosecutors without detention on 17 September 2026; police say the victims are teachers (more than one, number withheld) and no students at the school, the teachers' union questions whether the full scope of harm was confirmed, and the union says the victim teachers have been in treatment for two months while teaching daily, were questioned without a lawyer and wrote the criminal complaint themselves","summary":"Jeju Dongbu police announced on 23 September 2026 that a student at a middle school on Jeju had been referred to prosecutors without detention on 17 September for violating the Sexual Violence Punishment Act (filming with a camera or similar device). The student's phone was voluntarily submitted by his guardians and forensically examined; the examination found footage relating to some teachers and some sexual composites, and the student admitted the conduct. News1 reported that the student was booked for filming a teacher's body with his phone during a school field trip around mid-July 2026, that police found the offending had continued for several months with multiple victims, and that police found no indication the composites had been distributed. Police told Seoul Shinmun that there are more victim teachers than the one first reported but declined to give a number because the teachers are still teaching at the school and disclosure could cause secondary harm; that some suggestive composites were found and material causing shame was included in the charges; and that, contrary to earlier suggestions, no student at the school was a victim. The Jeju teachers' union questioned whether the full extent of the harm had been established, saying the victim teachers had themselves seen footage involving several students yet no compulsory seizure was made; police say the voluntary submission and forensic examination secured the necessary evidence. According to the union, two months after the case came to light the victim teachers are receiving hospital treatment and counselling while going to school every day; on 14 September they gave police victim statements without a lawyer and on 17 September attended the regional teachers' rights protection committee without support from the education office or the school safety mutual-aid association; they wrote and filed the criminal complaint themselves; substitute teachers are hard to find so they have used only three or four of 16 days of special leave; and the mutual-aid association proposes 3.3 million won in legal fees for all victim teachers together, payable after the case ends. The victim teachers requested meetings with the superintendent on 1 and 3 September without success, and a third written request was sent to the superintendent's office on 21 September; the union is running a petition until 2 October for severe punishment and support. Police said they had told the education office that the teachers need support such as treatment costs, rest and sick leave.","incidentDate":"2026-07-01","incidentKind":"ongoing_experience","incidentDatePrecision":"month","exposurePattern":"unknown","reportedDate":"2026-09-23","aiSystem":"AI image-compositing ('deepfake') tool; not named in the police announcement","aiProduct":"Unidentified image tool","severity":"high","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["exploitation_or_abuse","psychological_distress","professional_harm"],"harmOutcomeSummary":"Teachers were secretly filmed by a student and some footage was turned into sexual deepfake composites; two months on, per the union, the victim teachers are in hospital treatment and counselling while continuing to teach at the same school and handled the police statement, the protection committee and the criminal complaint without legal or institutional support. The filming and composites are admitted per police; the teachers' situation is the union's account.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":2,"affectedCountStatus":"documented_minimum","affectedCountEvidence":"Police told Seoul Shinmun there are victim teachers beyond the one first reported but withheld the number, and told News1 the examination confirmed multiple victims; two is the documented minimum. No student victims at the school per police. Documented minimum 2.","victimAgeRange":"adult","jurisdiction":"KR","platformType":"other","outcomeType":"criminal_charges","outcomeStatus":"ongoing","primarySourceUrl":"https://www.seoul.co.kr/news/society/accident/2026/09/23/20260923500094","primarySourceLabel":"Seoul Shinmun, 23 September 2026: police say there were no student victims; middle-school student who secretly filmed teachers and made deepfakes referred to prosecutors without detention","firstPublishedAt":"2026-09-23T03:45:58.820151+00:00","updatedAt":"2026-09-30T01:17:41.8046+00:00","scopeVersion":"facts-v3","tags":["deepfake","south-korea","jeju","teachers","school","illegal-filming","prosecutor-referral","depicted-or-impersonated"]},{"id":"2026-boca-raton-lyft-ai-damage-claim","caseFacts":{"claims":[{"id":"c1","status":"corroborated","evidence":[{"locator":"'She goes, \"Dad, that's AI, you could see the Gemini logo on the bottom right-hand side of one of the photos\"'; the $75 damage fee and the beach trip.","relation":"supports","source_id":"s1"},{"locator":"'A Boca Raton father says a Lyft driver used artificial intelligence to fabricate photos of car damage and charge his daughters a $75 damage fee'; 'that's the AI hologram on the photo'.","relation":"supports","source_id":"s2"},{"locator":"'his daughter Ella spotted an AI logo in the bottom righthand corner of the photo'; the $75 charge and the emergency-room context.","relation":"supports","source_id":"s3"}],"assertion":"On 16 May 2026, after a Lyft ride from Red Reef Park in Boca Raton, Bert Gor was charged a $75 damage fee; the driver's proof photo carried a Google Gemini watermark that his daughter spotted, showing it was AI-generated.","causal_attribution":"Reported identically in three independent interviews with the family."},{"id":"c2","status":"corroborated","evidence":[{"locator":"Lyft statement: 'We have reviewed the rider's concerns, offered reimbursement, and permanently removed the driver from the platform.'","relation":"supports","source_id":"s2"},{"locator":"Lyft statement to ABC News: 'offered reimbursement, and permanently removed the driver from the platform.'","relation":"supports","source_id":"s3"},{"locator":"'the Lyft rep agreed that the images were created using AI, apologized, and blocked the driver'; WPBF statement: 'offered reimbursement, unpaired the rider and driver, and addressed the matter directly with the driver.'","relation":"supports","source_id":"s1"}],"assertion":"Lyft agreed the damage photo was AI-generated, apologised, reimbursed the $75 charge and removed the driver from its platform.","causal_attribution":"Lyft statements to three outlets; the WPBF wording (unpaired/addressed) differs slightly from the WPTV/ABC wording (permanently removed)."}],"effects":[{"label":"father wrongly charged a $75 damage fee and daughters falsely accused of vandalism, on an AI-generated photo; Lyft reversed the charge and removed the driver","claim_id":"c1","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.wpbf.com/article/lyft-driver-ai-falsely-accuse-damaging-car/71341981","kind":"news_report","access":"read","language":"en","translation_note":"Read in English on 2026-09-17 via the Internet Archive capture 20260915020212 (WPBF live returns HTTP 451 to EEA requests); datePublished 2026-05-18. Original WPBF 25 interview with Bert Gor.","independence_group":"wpbf"},{"id":"s2","url":"https://www.wptv.com/news/local-news/our-community/boca-raton-delray-beach/boca-raton-father-catches-lyft-driver-using-ai-to-fake-car-damage-driver-removed","kind":"news_report","access":"read","language":"en","translation_note":"Read in English on 2026-09-17 (datePublished 2026-05-21). WPTV interview with Gor; notes parts of the WPTV story were AI-assisted and staff-reviewed.","independence_group":"wptv"},{"id":"s3","url":"https://abcnews.com/GMA/News/father-daughter-speak-after-lyft-driver-accused-ai/story?id=133144373","kind":"news_report","access":"read","language":"en","translation_note":"Read in English on 2026-09-17 (datePublished 2026-05-20). ABC Good Morning America interview with Bert and Ella Gor; carries the Lyft statement that the driver was permanently removed.","independence_group":"abc"}],"version":1,"ai_roles":["others_use"],"contexts":["everyday_life","finance"],"unknowns":["The driver is not named and faced no reported police or legal action.","Whether Lyft changed its damage-dispute review process is not established.","How many other riders were affected by similar fabricated claims is not quantified beyond Gor's Facebook responses."],"geography":{"basis":"Event: a Lyft ride ending at the Gor family home in Boca Raton, Florida (WPBF, WPTV). Affected people are Boca Raton residents. No court.","court_countries":[],"event_countries":["US"],"affected_person_countries":["US"]},"publication":{"basis":"Published under the 2026-09-15 charter as an ordinary adverse experience involving an AI-fabricated artifact: a named father and his teenage daughters were wrongly charged and accused on the strength of an AI-generated damage photo, and the company reversed the charge and removed the driver. Low severity; three independent interview-based chains.","reviewed_on":"2026-09-17"},"ai_involvement":{"basis":"The family identified a Google Gemini watermark on the damage photo, and Lyft agreed the image was AI-generated. The AI was used by a third party (the driver) to fabricate evidence; it did not communicate with, depict or make an autonomous claim about the affected family, so the person relation is recorded as unknown.","status":"supported"},"person_relations":["unknown"]},"name":"Boca Raton, Florida: a Lyft driver used an AI-generated photo to charge a father a false $75 damage fee after his daughters' ride","summary":"On 16 May 2026 two teenage sisters (14 and 15) and a friend took a Lyft home from Red Reef Park in Boca Raton, Florida. Their father, Bert Gor, was then charged a $75 damage fee after the driver told Lyft's customer service the girls had left a mess. When Gor asked for proof, the photo he was sent carried a Google Gemini watermark that one daughter recognised, showing the image had been AI-generated. Lyft agreed the image was fake, reimbursed the charge and permanently removed the driver from its platform.","incidentDate":"2026-05-16","incidentKind":"single_event","incidentDatePrecision":"day","exposurePattern":"single_interaction","reportedDate":"2026-05-18","aiSystem":"AI image generation (a Google Gemini-watermarked image used by a Lyft driver to fabricate a damage-claim photo)","aiProduct":"Gemini (reported)","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["financial_loss","reputational_harm"],"harmOutcomeSummary":"A father was charged a fabricated $75 damage fee and his two teenage daughters were falsely accused of vandalising a rideshare vehicle, on the strength of an AI-generated photo; Lyft reversed the charge and removed the driver after the family identified the AI watermark.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":3,"affectedCountStatus":"documented_minimum","affectedCountEvidence":"Three identified affected people: Bert Gor (charged the $75 fee) and his two teenage daughters (falsely accused of leaving a mess). A friend also rode but is not reported as charged or accused. Documented minimum of three.","victimAgeRange":"minor","jurisdiction":"US","platformType":"other","outcomeType":"internal_action","outcomeStatus":"resolved","primarySourceUrl":"https://www.wpbf.com/article/lyft-driver-ai-falsely-accuse-damaging-car/71341981","primarySourceLabel":"WPBF 25 News (West Palm Beach), 18 May 2026: Boca Raton father says Lyft driver used AI to fabricate a damage photo (interview with Bert Gor; Lyft statement)","firstPublishedAt":"2026-09-17T03:27:40.160475+00:00","updatedAt":"2026-09-30T01:17:29.168447+00:00","scopeVersion":"facts-v3","tags":["ai-generated-image","deepfake-evidence","rideshare","lyft","gemini","fabricated-evidence","united-states","florida","consumer","everyday-life"]},{"id":"2026-christi-hill-grok-false-officer-identification","caseFacts":{"claims":[{"id":"c1","status":"corroborated","evidence":[{"locator":"'Grok has falsely named Hill and the other officer as the \"primary officers shown\" in the bodycam footage released by Hampshire police'; caption: 'Screengrab of a post by Grok referencing Christy Hill being incorrectly named'.","relation":"supports","source_id":"s1"},{"locator":"'she was wrongly identified by social media users and artificial intelligence (AI) chatbot Grok'; Hill 'criticised Grok ... for the \"chaos\" it caused by misidentifying her'.","relation":"supports","source_id":"s2"}],"assertion":"Grok falsely named Hill and another officer as the 'primary officers shown' in the bodycam footage of Nowak's arrest; the Guardian carries a screengrab of a Grok post incorrectly naming her.","causal_attribution":"The Guardian documents the Grok post with a screengrab; BBC Verify's interview independently carries her account of the Grok misidentification. What Grok answered to particular queries beyond the screengrabbed post is not documented."},{"id":"c2","status":"corroborated","evidence":[{"locator":"'Hill served as an officer in Portsmouth for 12 years before leaving the force in April 2024 – 20 months before the murder took place'; her statement: 'a national police bravery award media release... is being repeatedly shared and misattributed to this case'.","relation":"supports","source_id":"s1"},{"locator":"'Hill said she left Hampshire Constabulary in April 2024 - more than a year before Nowak was fatally stabbed in December 2025'.","relation":"supports","source_id":"s2"}],"assertion":"The identification is factually impossible: Hill left Hampshire Constabulary in April 2024, more than a year before the December 2025 murder; the confusion stems from a national police bravery award photo of her and a former colleague being misattributed.","causal_attribution":"Her account, uncontested in the inspected sources; the bravery-award photo mechanism is her explanation, attributed."},{"id":"c3","status":"corroborated","evidence":[{"locator":"'forced to flee to a safe space'; Hampshire: 'A police officer unrelated to this case has been misidentified online and subject to death threats.'","relation":"supports","source_id":"s1"},{"locator":"'scared for her safety and has been hiding in a safe house'; 'leading to \"threats of violence\"'; video caption: \"received 'endless threats'\".","relation":"supports","source_id":"s2"}],"assertion":"After the false posts, Hill received 'endless threats' and 'threats of violence' and is hiding in a safe house; Hampshire confirmed an unrelated officer was misidentified online and subjected to death threats.","causal_attribution":"Her account to two independent outlets plus the constabulary's own confirmation."},{"id":"c4","status":"corroborated","evidence":[{"locator":"'The home secretary, Shabana Mahmood, said on Tuesday that a male officer had been misidentified and that he had had to move out of his home.'","relation":"supports","source_id":"s1"},{"locator":"'BBC Verify has also obtained evidence that current police officer Tristan Parsons, who was also wrongly accused online of being involved, was not in the country at the time of Nowak's murder.'","relation":"supports","source_id":"s2"}],"assertion":"A second officer was wrongly accused in the same wave: the Home Secretary said a male officer had been misidentified and had to move out of his home; BBC Verify obtained evidence that serving officer Tristan Parsons was not in the country at the time of the murder.","causal_attribution":"The Home Secretary's account covers the displacement; BBC Verify's evidence covers Parsons's non-involvement. Whether Mahmood meant Parsons specifically is not stated; the two accounts are recorded separately."},{"id":"c5","status":"reported","evidence":[{"locator":"'Musk... posting on his X platform during the trial that he would fund a private prosecution of the officers involved.'","relation":"supports","source_id":"s1"},{"locator":"Starmer passage: 'Musk, again, has been interfering in our politics in the last few days, trying to whip up division'.","relation":"supports","source_id":"s2"}],"assertion":"Elon Musk posted during the trial that he would fund a private prosecution of the officers involved; Prime Minister Keir Starmer accused Musk of trying to 'whip up division' over the case.","causal_attribution":"Context for the amplification climate; not itself an AI claim."}],"effects":[{"label":"false AI identification of a former officer in a charged murder case; threats drove her into hiding","claim_id":"c3","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.theguardian.com/uk-news/2026/jun/03/former-officer-hampshire-hiding-after-being-falsely-linked-henry-nowak-arrest","kind":"news_report","access":"read","language":"en","translation_note":"Read in English on 2026-09-15 (datePublished 2026-06-03T16:28Z). Includes Hill's statement, the Grok screengrab reference, the Hampshire statement and the Home Secretary's account.","independence_group":"guardian"},{"id":"s2","url":"https://www.bbc.com/news/articles/cg7pl7zj024o","kind":"news_report","access":"read","language":"en","translation_note":"Read in English on 2026-09-15 (datePublished 2026-06-04T15:50Z; BBC Verify, Devlin & Nguyen, with André Rhoden-Paul).","independence_group":"bbc"}],"version":1,"ai_roles":["unknown"],"contexts":["justice","everyday_life"],"unknowns":["What Grok answered in sessions other than the screengrabbed post, and whether xAI corrected or constrained the false identification, is not reported.","The date the false posts began (during the trial) is not precisely stated; Hill first saw them around 03:00 BST on 2 June 2026.","Whether any platform removed the misidentifying content is not reported.","The jury inquest into Nowak's death is expected next year; its findings are pending and concern the police response, not the misidentification."],"geography":{"basis":"Event: the online spread about the Hampshire (Southampton) case and Hill's concealment in the UK. Affected persons: both officers are UK-based (Parsons was abroad at the time of the murder, not at the time of the misidentification). Court country: the Digwa murder trial and the expected inquest are in England; no proceeding by or against the misidentified officers is reported.","court_countries":["GB"],"event_countries":["GB"],"affected_person_countries":["GB"]},"publication":{"basis":"Published under the 2026-09-15 charter as a consequential false claim about identified people (made_claim_about): two officers wrongly named by AI platforms in a charged case, documented by two independent chains including the outlet's screengrab evidence and the constabulary's confirmation. Hill spoke publicly; Parsons is named by BBC Verify in a vindicating context. The Nowak case background is limited to what the claims require.","reviewed_on":"2026-09-15"},"ai_involvement":{"basis":"The Guardian carries a screengrab of a Grok post incorrectly naming Hill and reports in its own voice that Grok falsely named her and the other officer as the 'primary officers shown'; Hill's own statement names 'AI platforms such as Grok'; BBC Verify's interview confirms the Grok misidentification. xAI was approached for comment by both outlets.","status":"supported"},"person_relations":["made_claim_about"]},"name":"Grok falsely named former officer Christi Hill as an arresting officer in the Henry Nowak case; she is in hiding after threats","summary":"In early June 2026, social media posts and the AI chatbot Grok falsely identified former Hampshire police constable Christi Hill as one of the 'primary officers shown' in bodycam footage of the December 2025 arrest of Henry Nowak, who died after being stabbed. Hill had left the force in April 2024 — more than a year before the murder — and says the misidentification stems from a national police bravery award photo being misattributed. After 'endless threats' and 'threats of violence' she is hiding in a safe house. BBC Verify separately obtained evidence that serving officer Tristan Parsons, also wrongly accused online, was not in the country at the time; the Home Secretary said a misidentified male officer had to move out of his home. Hampshire Constabulary confirmed an unrelated officer was misidentified and received death threats.","incidentDate":"2026-06-02","incidentEndDate":"2026-06-04","incidentKind":"bounded_series","incidentDatePrecision":"range","exposurePattern":"repeated_interactions","reportedDate":"2026-06-03","aiSystem":"Grok (and other AI platforms, unnamed)","aiProduct":"Grok","aiCompany":"xAI","severity":"high","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["psychological_distress","reputational_harm","loss_of_autonomy"],"harmOutcomeSummary":"A factually impossible identification by AI platforms including Grok put a former officer into hiding after death threats; a serving officer wrongly accused in the same wave had to move out of his home, per the Home Secretary.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":2,"affectedCountStatus":"exact","affectedCountEvidence":"Two documented harmed people: Hill (Guardian interview and statement; BBC Verify interview) and the male officer identified by BBC Verify as Tristan Parsons (BBC's evidence he was abroad; the Home Secretary's account that a male officer had to move home). No further misidentified officers are reported in the inspected sources.","victimAgeRange":"adult","jurisdiction":"GB","platformType":"chatbot","outcomeType":"media_coverage","outcomeStatus":"ongoing","primarySourceUrl":"https://www.theguardian.com/uk-news/2026/jun/03/former-officer-hampshire-hiding-after-being-falsely-linked-henry-nowak-arrest","primarySourceLabel":"The Guardian (3 Jun 2026) — Hill's statement, the Grok screengrab, Hampshire and Mahmood accounts","firstPublishedAt":"2026-09-15T09:25:27.124866+00:00","updatedAt":"2026-09-30T01:17:32.056946+00:00","scopeVersion":"facts-v3","tags":["made-claim-about","false-identification","grok","police","threats","uk"]},{"id":"2026-bengaluru-gemini-triple-murder","metadataReview":{"version":1,"geography":{"basis":"Geographic fields reviewed against the cited retained dossier passages. This is a metadata review, not a new source or causal review. Residence, event setting and court are recorded separately. Affected-person countries refer to reported location or residence, not citizenship or a complete residence history.","evidence":[{"kind":"existing_record","locator":"All killed in Bengaluru's KR Puram area","location":"victim_description","supports":["event_countries","affected_person_countries"],"countries":["IN"]}],"court_countries":[],"event_countries":["IN"],"affected_person_countries":["IN"]},"unresolved":["No court country established in this review. Regulatory authorities and police locations are not automatically courts.","Source languages remain unreviewed; no source body was newly inspected for this record during this metadata pass."],"reviewed_on":"2026-09-15","source_reviews":[]},"name":"Bengaluru KR Puram Triple Murder (Kenneth — Google Gemini Planning)","summary":"On 22 June 2026 in Bengaluru's KR Puram area, 25-year-old J. Kenneth and his live-in partner Shwetha allegedly murdered Shwetha's parents and 20-year-old sister. Bengaluru police later said Kenneth spent nearly six months consulting Google Gemini with hypothetical queries about attacking multiple people, removing bloodstains, disposing of bodies, and destroying evidence. Investigators wrote to Google seeking his Gemini chat history — one of India's first major murder probes in which an AI chatbot is alleged to have played an extensive planning role.","incidentDate":"2026-06-22","incidentKind":"single_event","incidentDatePrecision":"day","exposurePattern":"repeated_interactions","reportedDate":"2026-07-22","aiSystem":"Google Gemini","aiProduct":"Gemini","aiCompany":"Google / Alphabet","severity":"critical","verificationStatus":"credible","harmCategories":["third_party_harm_facilitation","method_provision"],"harmOutcomes":["death","harm_to_others","legal_harm","loss_of_liberty"],"harmOutcomeSummary":"Three people were killed and two participants were arrested after police said one had spent about six months consulting Gemini about attacks, blood removal, body disposal and destruction of evidence.","frameworkFacets":["P1d","P5d"],"pathwayDirection":"outward","causationStatus":"alleged","participantUsersAffectedMin":1,"otherPeopleHarmedMin":3,"affectedCountStatus":"exact","affectedCountEvidence":"Police reporting identifies one accused person as the AI participant and names three homicide victims. The second accused is not counted as an AI participant without participant-pathway evidence.","victimAgeRange":"adult","jurisdiction":"IN-KA","platformType":"assistant","outcomeType":"criminal_charges","outcomeStatus":"ongoing","primarySourceUrl":"https://www.indiatoday.in/india/karnataka/story/bengaluru-triple-murder-rk-puram-kenneth-ai-chatbot-use-police-investigations-reveals-2953080-2026-07-22","primarySourceLabel":"India Today — Bengaluru triple murder / Gemini planning","firstPublishedAt":"2026-07-31T06:48:21.689201+00:00","updatedAt":"2026-09-30T01:17:27.042071+00:00","scopeVersion":"participant-v2","tags":["criminal","adult","murder","gemini","google","third-party-harm","method-provision","india","in-ka"]},{"id":"2026-asato-v-xai-grok","caseFacts":{"claims":[{"id":"2026-09-14-pilot-o12-c1","status":"reported","evidence":[{"locator":"Background section.","relation":"supports","source_id":"2026-09-14-pilot-s16"}],"assertion":"A claimant's solicitors state that Grok was used to produce non-consensual sexualized images of her.","causal_attribution":"Attributed in the account; independent causation has not been established."},{"id":"2026-09-14-pilot-o12-c2","status":"reported","evidence":[{"locator":"Filing announcement paragraph.","relation":"supports","source_id":"2026-09-14-pilot-s16"}],"assertion":"The solicitors announce a claim filed in the High Court in England on June 3, 2026.","causal_attribution":"Procedural statement by the claimant’s representatives; no finding on liability is asserted."}],"effects":[{"label":"nonconsensual_sexualized_imagery","claim_id":"2026-09-14-pilot-o12-c1","direction":"negative"}],"sources":[{"id":"2026-09-14-pilot-s16","url":"https://awo.agency/articles/first-claim-in-the-uk-against-groks-nonconsensual-deepfakes/","kind":"legal_representative_statement","access":"read","language":"en","translation_note":"Read in English; no translation needed.","independence_group":"2026-09-14-pilot-g13"}],"version":1,"ai_roles":["others_use"],"contexts":["privacy","public_life","deepfakes","non-consensual-imagery"],"unknowns":["The probe did not inspect the court file or independently adjudicate liability."],"geography":{"basis":"Affected-person country from the historical record and solicitor's account; court country explicitly stated. Location of online image generation unknown.","court_countries":["GB"],"event_countries":[],"affected_person_countries":["GB"]},"publication":{"basis":"Restored a previously published case under its original ID after source, duplicate and sensitivity review. Allegations remain attributed to the claimant’s solicitors.","reviewed_on":"2026-09-15"},"ai_involvement":{"basis":"The account or representative reports use of an AI system; independent verification was not performed.","status":"reported"},"person_relations":["depicted_or_impersonated"]},"name":"Jess Asato’s solicitors report non-consensual Grok imagery and a High Court claim","summary":"Jess Asato’s solicitors report that Grok was used to make non-consensual sexualized images of her in January 2026. They announced a High Court claim against xAI on 3 June. These are the claimant’s representatives’ accounts; liability has not been established by the evidence reviewed here.","incidentDate":"2026-01-01","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"unknown","reportedDate":"2026-06-03","aiSystem":"Grok","aiProduct":"Grok","aiCompany":"xAI","severity":"high","verificationStatus":"unverified","harmCategories":[],"harmOutcomes":["exploitation_or_abuse"],"harmOutcomeSummary":"A claimant's solicitors state that Grok was used to produce non-consensual sexualized images of her.","frameworkFacets":[],"causationStatus":"alleged","participantUsersAffectedMin":0,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"One person is described as harmed in this account; wider affected populations are not counted.","victimAgeRange":"adult","outcomeType":"lawsuit_filed","outcomeStatus":"unknown","primarySourceUrl":"https://awo.agency/articles/first-claim-in-the-uk-against-groks-nonconsensual-deepfakes/","primarySourceLabel":"AWO, representing the claimant","firstPublishedAt":"2026-06-14T11:10:33.428074+00:00","updatedAt":"2026-09-30T01:17:24.769959+00:00","scopeVersion":"facts-v3","tags":["privacy","public_life","deepfakes","non-consensual-imagery"]}]}