{"meta":{"exportedAt":"2026-10-23T09:23:53.724Z","formatVersion":2,"selection":{"q":"investigation","system":"","harm":"","context":"","country":"","role":"","relation":"contextual","evidence":"","year":"","response":"regulatory","severity":"","verification":"","view":"incidents","sort":"added"},"totalIncidents":8,"coverage":{"cases":8,"countries":5,"languages":7,"unknownLocation":2,"locationPending":0,"unknownLanguage":0,"unknownDate":0,"lawsuits":0,"regulatory":8,"minors":1,"coreRelations":2,"contextualRelations":8,"mixedRelations":2,"unknownRelations":0,"relationPending":0,"relationUnknown":0},"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":"2023-victoria-child-protection-worker-entered-family-information-into-chatgpt-for-court-report-regulator-finds","caseFacts":{"claims":[{"id":"c1","status":"documented","evidence":[{"locator":"Report: 'The Deputy Commissioner therefore found that DFFH contravened both IPP 3.1 and IPP 4.1 and determined that the contraventions were “serious”'; 'Issue a direction to Child Protection staff setting out that they are not to use any web-based or external Application Programming Interface (API)-based GenAI text tools (such as ChatGPT) as part of their official duties'; 'This action must be implemented by 5 November 2024 and maintained until 5 November 2026'.","relation":"supports","source_id":"s1"},{"locator":"'The regulator issued orders to the department, including blocking IP and domains for a range of generative AI websites, including ChatGPT, Meta AI, Gemini and Copilot.'","relation":"supports","source_id":"s2"}],"assertion":"OVIC's investigation report, published in September 2024, found that DFFH contravened Information Privacy Principles 3.1 and 4.1 over a child protection worker's use of ChatGPT, that the contraventions were serious, and issued a compliance notice requiring DFFH to direct child protection staff not to use web-based generative AI text tools and to block them until 5 November 2026.","causal_attribution":"The regulator's own report and compliance notice establish its findings and orders."},{"id":"c2","status":"documented","evidence":[{"locator":"Compliance notice: 'In or around September 2023, a Child Protection worker (CPW1) of the Organisation used'; 'ChatGPT, a web-based generative Artificial Intelligence (GenAI) tool, when drafting a Protection'. Report: 'The report had been submitted to the Children’s Court for a case concerning a young child whose parents had been charged in relation to sexual offences'; 'CPW1 admitted that they had used ChatGPT in drafting the report, but denied inputting personal information as part of this'.","relation":"supports","source_id":"s1"}],"assertion":"OVIC found that in or around September 2023 a DFFH child protection worker used ChatGPT when drafting a Protection Application Report that was submitted to the Children's Court in a case about a young child whose parents had been charged in relation to sexual offences; the worker admitted to DFFH using ChatGPT for the report.","causal_attribution":"OVIC's finding, based on indicators in the report and the worker's admission to DFFH."},{"id":"c3","status":"documented","evidence":[{"locator":"Report: 'OVIC found, on the balance of probabilities, that CPW1 input personal information, including names'; 'By entering personal and sensitive information about the mother, father, carer, and child into ChatGPT, CPW1 also disclosed this information to OpenAI (the company which operates ChatGPT)'; DFFH response: 'it is not possible to definitively confirm that personal information has ever been entered into ChatGPT by the Child Protection Practitioner who no longer works in the department'.","relation":"supports","source_id":"s1"},{"locator":"'Ovic found that, on the balance of probabilities, the worker had entered in personal information.'","relation":"supports","source_id":"s2"}],"assertion":"OVIC found on the balance of probabilities that the worker entered personal information, including names, about the child, mother, father and carer into ChatGPT, disclosing it to OpenAI; the worker denied entering personal information, and DFFH says this cannot be confirmed definitively.","causal_attribution":"A regulator's finding on the balance of probabilities, contested by the worker's denial; organisations cannot see what staff typed into ChatGPT."},{"id":"c4","status":"documented","evidence":[{"locator":"Report: 'First, the PA Report referenced how the child’s father used a particular toy, a doll, for sexual purposes.'; 'The report later referenced the same toy as a notable strength, in that the parents had provided the child with “age-appropriate toys”'; 'ChatGPT failed to properly understand the relevant context, and generated inappropriate and inaccurate content as a result'; 'these deficiencies in the report did not ultimately change the decision-making of either Child Protection or the Court in relation to the child'.","relation":"supports","source_id":"s1"},{"locator":"'But it also said the report did not change the decision-making of either the child protection agency or the court.'","relation":"supports","source_id":"s2"}],"assertion":"OVIC's report attributes inaccurate content in the report to ChatGPT: having described a doll the father had reportedly used for sexual purposes, the report later cited the same toy as a parental strength ('age-appropriate toys'), which downplayed the risk to the child; the deficiencies did not change the decisions of Child Protection or the court.","causal_attribution":"OVIC attributes the doll passage to ChatGPT from the content itself and the worker's admission of ChatGPT use."},{"id":"c5","status":"reported","evidence":[{"locator":"Report: 'DFFH arranged with the Court for the PA Report to be withdrawn and a replacement one submitted. It also notified affected parties about the incident'; 'In the present case, the mother, father, and carer variously described feelings of anxiety and being overwhelmed when informed about their information being disclosed'.","relation":"supports","source_id":"s1"}],"assertion":"According to OVIC's report, DFFH arranged with the court for the report to be withdrawn and replaced and notified the affected parties; the mother, father and carer described feelings of anxiety and being overwhelmed when told their information had been disclosed.","causal_attribution":"DFFH's account to OVIC; the family members' reactions are relayed by the regulator, not in their own words."},{"id":"c6","status":"documented","evidence":[{"locator":"DFFH response, 18 September 2024: 'I accept that there was an unauthorized use of ChatGPT by a Child Protection Practitioner'; 'when developing a Protection Application Report for court'; 'the Child Protection Practitioner who no longer works in the department'.","relation":"supports","source_id":"s1"},{"locator":"'In response, the department said it accepted the findings and would work to implement the orders. It said it no longer employed the worker.'","relation":"supports","source_id":"s2"}],"assertion":"DFFH's Secretary accepted that a child protection practitioner had used ChatGPT without authorisation for a court report and that the department breached the two privacy principles, and the department said it no longer employs the worker.","causal_attribution":"DFFH's own letter, annexed to the report."}],"effects":[{"label":"personal information about a child, the mother, father and carer was entered into ChatGPT and disclosed to OpenAI, regulator finds","claim_id":"c3","direction":"negative"},{"label":"ChatGPT-generated content in a Children's Court report downplayed the risk to the child; the report was withdrawn and replaced","claim_id":"c4","direction":"negative"},{"label":"the mother, father and carer described anxiety and being overwhelmed when told of the disclosure","claim_id":"c5","direction":"negative"}],"sources":[{"id":"s1","url":"https://ovic.vic.gov.au/wp-content/uploads/2024/11/DFFH-ChatGPT-investigation-report-20240924-Re-upload.pdf","kind":"official_record","access":"read","language":"en","translation_note":"OVIC investigation report under s8C(2)(e) of the Privacy and Data Protection Act 2014 (Vic), September 2024, with the compliance notice (Annexure A) and DFFH's response letter of 18 September 2024 (Annexure B); the PDF (re-uploaded November 2024) was read in full on 2026-10-22 from ovic.vic.gov.au (HTTP 200); the report page gives 'Last updated 24 September 2024'.","independence_group":"ovic-dffh-chatgpt-investigation"},{"id":"s2","url":"https://www.theguardian.com/australia-news/2024/sep/26/victoria-child-protection-chat-gpt-ban-ovic-report-ntwnfb","kind":"news","access":"read","language":"en","translation_note":"The Guardian (Australia), 26 September 2024, read in full on 2026-10-22 (HTTP 200). Its account comes from OVIC's report, so it shares that group. Its statement that the worker may have used ChatGPT in 100 cases does not match the report, which attributes the 100 cases to the worker's broader work unit; that statement is not relied on.","independence_group":"ovic-dffh-chatgpt-investigation"}],"version":1,"ai_roles":["institutional_use"],"contexts":["privacy","public_services","justice"],"unknowns":["Exactly what text the worker entered into ChatGPT; neither DFFH nor OVIC could see the prompts, and the input of personal information rests on a balance-of-probabilities finding.","The exact date of the ChatGPT use (the compliance notice says 'in or around September 2023') and of the court hearing.","Whether OpenAI retained or used the information entered.","Which people were affected in the 100 other cases with indicators of ChatGPT use, and whether their information was entered.","How DFFH has complied with the compliance notice since its first monitoring reports."],"geography":{"basis":"OVIC is Victoria's information regulator, DFFH is a Victorian government department, and the report says the affected child protection case went before the Children's Court, a Victorian court. The family's case was handled by the department in Victoria.","court_countries":["AU"],"event_countries":["AU"],"affected_person_countries":["AU"]},"publication":{"basis":"Published under the 2026-10-06 privacy clarification: a regulator's investigation report finds that a child protection worker entered a family's personal information into ChatGPT, without authorisation, and that ChatGPT-generated text minimising the risk to a child went into a court report. Relations: received_personal_information and made_claim_about. The case adds a statutory child protection setting, a court report and a compliance notice banning generative AI tools.","reviewed_on":"2026-10-22"},"ai_involvement":{"basis":"OVIC's report finds that ChatGPT received personal information about the child, mother, father and carer that the worker typed in, which disclosed it to OpenAI, and that ChatGPT generated report text inaccurately describing the parents' care of the child (the doll passage), which went into a report submitted to the Children's Court and downplayed the risk to the child. The finding rests on indicators in the report text, the worker's admission of ChatGPT use and colleagues' accounts; the input of personal information was found on the balance of probabilities and the worker denied it.","status":"supported"},"person_relations":["received_personal_information","made_claim_about"]},"name":"Victorian child protection worker put a family's information into ChatGPT to draft a court report that minimised risk to the child, regulator finds","summary":"Victoria's Information Commissioner (OVIC) found in a report published on 24 September 2024 that a child protection worker at the Department of Families, Fairness and Housing (DFFH) used ChatGPT in or around September 2023 to draft a Protection Application Report submitted to the Children's Court about a young child. On the balance of probabilities, OVIC found the worker entered personal information, including names, about the child, mother, father and carer, disclosing it to OpenAI; the worker admitted using ChatGPT but denied entering personal information. OVIC found generated content in the report described a doll that the father had reportedly used for sexual purposes as a parental strength ('age-appropriate toys'), downplaying the risk to the child, although the court's and the department's decisions did not change. The report was withdrawn and replaced, and the mother, father and carer described anxiety when told of the disclosure. OVIC found serious breaches of two privacy principles and issued a compliance notice requiring DFFH to ban and block generative AI text tools for child protection staff until November 2026; DFFH accepted the breach findings.","incidentDate":"2023-09-01","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"single_interaction","reportedDate":"2024-09-24","aiSystem":"ChatGPT (OpenAI), used through its website by a child protection worker of Victoria's Department of Families, Fairness and Housing to draft a court report, without departmental authorisation (per OVIC's investigation report)","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["other_material_harm","psychological_distress"],"harmOutcomeSummary":"OVIC found that personal information about a child, the mother, father and carer was entered into ChatGPT and disclosed to OpenAI, and that ChatGPT-generated content in a court report downplayed the risk to the child; the mother, father and carer described anxiety when told, according to OVIC's report.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":4,"affectedCountStatus":"documented_minimum","affectedCountEvidence":"OVIC's report finds that personal and sensitive information about the mother, father, carer and child in this case was entered into ChatGPT (four people), and records that the mother, father and carer described anxiety when informed. OVIC says personal information had likely been entered in other cases too; those people are not counted. The worker, the court and DFFH staff are not counted.","victimAgeRange":"unknown","jurisdiction":"AU-VIC","platformType":"chatbot","outcomeType":"regulatory_action","outcomeStatus":"resolved","primarySourceUrl":"https://ovic.vic.gov.au/wp-content/uploads/2024/11/DFFH-ChatGPT-investigation-report-20240924-Re-upload.pdf","primarySourceLabel":"Office of the Victorian Information Commissioner: Investigation into the use of ChatGPT by a Child Protection worker (report, 24 Sep 2024)","firstPublishedAt":"2026-10-22T03:03:14.588614+00:00","updatedAt":"2026-10-22T03:03:14.588614+00:00","scopeVersion":"facts-v3","tags":["chatgpt","child-protection","court-report","privacy-regulator","unauthorised-ai-use","personal-information","victoria"]},{"id":"2026-caine-prize-winning-story-pangram-ai-detector-accusation-prize-review","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'The central evidence of Vara’s article, “Yet Another Literary Prizewinner Seems to Be AI,” is a scan by Pangram, an AI-detection tool, which flags almost the entire story as AI-generated, with the epigraph as a notable exception.'","relation":"supports","source_id":"s2"},{"locator":"'a recent article in The Atlantic said AI-detection software Pangram had shown his work about a boy whose family falls on hard times was entirely AI generated'; 'had been published on Sunday'","relation":"supports","source_id":"s1"}],"assertion":"The Atlantic published an article on Sunday 11 October 2026 whose central evidence was a scan by the Pangram AI-text detection tool flagging almost all of the 2026 Caine Prize-winning story as AI-generated.","causal_attribution":"The journalist's article (one origin), read through Brittle Paper's and the BBC's accounts; The Atlantic page itself was not read. The Sunday date is the BBC's."},{"id":"c2","status":"reported","evidence":[{"locator":"'the prize is reviewing his winning story'; 'The Caine Prize said in a statement that it is its responsibility to look into such claims, and that it is taking a “human centred approach” that supports writers during the process. It added that the process has begun'; 'launched an investigation as soon as they learned of Vara’s inquiry'","relation":"supports","source_id":"s2"},{"locator":"'as part of an internal investigation for the Caine Prize'; 'would be sending in documentation about his various drafts and there were ongoing conversations with all the judges and The Shallow Tales Review'","relation":"supports","source_id":"s1"}],"assertion":"The Caine Prize began a review of the winning story, which its chair told the journalist was launched as soon as the prize learned of her inquiry: it said it was its responsibility to look into such claims, and its chair said the author would send documentation of his drafts and that the prize was talking to the judges and to the story's first publisher.","causal_attribution":"The prize's own statement and its chair's interview (one origin: the Caine Prize). No finding is reported."},{"id":"c3","status":"reported","evidence":[{"locator":"'they did have evidence of real human investigation of it before tainting the prize and tainting this winner'; 'against is it being used as a final arbiter in this case'","relation":"supports","source_id":"s1"}],"assertion":"The prize chair told the BBC that The Atlantic should have done more human investigation before tainting the prize and the winner, and that she was against a detector being the final arbiter in this case.","causal_attribution":"The chair's own words in a BBC interview (one origin)."},{"id":"c4","status":"reported","evidence":[{"locator":"'For the avoidance of doubt, I wrote The Blue Fighter,\" he posted on X'; 'had \"a concrete, documented history\" about the story's inspiration'","relation":"supports","source_id":"s1"},{"locator":"'literary agency has said it is confident the story is his own work'","relation":"supports","source_id":"s2"}],"assertion":"The author denied using AI to write the story and said he had a documented history of its inspiration and development; his literary agency said it was confident the story was his own work.","causal_attribution":"The author's own statement and his agency's (one side of the dispute); whether AI was used is unestablished."},{"id":"c5","status":"reported","evidence":[{"locator":"'Pangram's internal research showed its assessments were rarely wrong'; 'was \"not going to be 100% accurate\"'; 'she had also used her own judgment'","relation":"supports","source_id":"s1"},{"locator":"'Vara closes by saying she may well be wrong.'","relation":"supports","source_id":"s2"}],"assertion":"The journalist said Pangram's internal research shows its assessments are rarely wrong and that she also relied on her own judgement, while conceding that no AI-detection product is fully accurate.","causal_attribution":"The journalist's own account (one origin), reported by the BBC interview and Brittle Paper's summary of the article."}],"effects":[{"label":"a prize-winning story placed under the Caine Prize's review, with the author asked to document his drafts, after a published AI-detector accusation","claim_id":"c2","direction":"negative"},{"label":"a prize winner publicly accused of submitting an AI-written story on the strength of an AI-detector scan","claim_id":"c1","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.bbc.co.uk/news/articles/cm4g1rekpgnjo","kind":"news_report","access":"read","language":"en","translation_note":"Read in English by the research agent (an AI) on 2026-10-14 (live, HTTP 200; datePublished 13 October 2026). BBC's own Front Row and Newsday interviews with the prize chair, the journalist and the author, plus the author's X statement.","independence_group":"bbc-front-row"},{"id":"s2","url":"https://brittlepaper.com/2026/10/african-literary-community-pushes-back-on-the-atlantics-ai-claims-about-10000-caine-prize-winner-simbarashe-steyn-kundizeza/","kind":"news_report","access":"read","language":"en","translation_note":"Read in English by the research agent (an AI) on 2026-10-14 (live, HTTP 200; dated 13 October 2026). Summarises The Atlantic's article and quotes the Caine Prize statement; The Atlantic page itself returned HTTP 403.","independence_group":"brittle-paper"}],"version":1,"ai_roles":["others_use"],"contexts":["work"],"unknowns":["Whether generative AI was used to write the story; the author denies it and the prize has announced no finding.","The outcome of the Caine Prize review and whether the prize will be withdrawn.","The Atlantic's article was not read directly (HTTP 403); its methods are known through the BBC's and Brittle Paper's accounts.","Any personal or financial effect on the author beyond the public accusation and the review."],"geography":{"basis":"No source states where the accusation, the review or the author is located for this case. The prize is the Caine Prize for African Writing and the author is described as Zimbabwean, but nationality and the publisher's or prize's base are not used to locate the event.","court_countries":[],"event_countries":[],"affected_person_countries":[]},"publication":{"basis":"Published as a contextual case: an AI-text detector's output about a prize-winning story was the central evidence of a national magazine's accusation; the Caine Prize opened a review when it learned of the journalist's inquiry and asked the author for his drafts (BBC; Brittle Paper). AI authorship is disputed and unproven; the author's denial and the prize chair's objection to detector evidence are recorded. Low severity; the author is not named.","reviewed_on":"2026-10-14"},"ai_involvement":{"basis":"The output: a Pangram AI-text detector scan that flagged almost the entire winning story as AI-generated, which was the central evidence of The Atlantic's article (c1; Brittle Paper; BBC). The connection to the harm: the Caine Prize began a review when it learned of the journalist's inquiry, which rested on the scan, and continued it after the article of the story and asked the author for documentation of his drafts (c2), and its chair said the article had tainted the winner (c3). The detector's output is a claim about the author's text; whether it is correct is disputed, as the author denies using AI (c4) and the prize chair rejects a detector as the final arbiter. The Atlantic page itself was not read, so the scan is known through the BBC's and Brittle Paper's accounts (made_claim_about).","status":"reported"},"person_relations":["made_claim_about"]},"name":"Caine Prize reviews its 2026 winning story after a Pangram AI-detector scan cited by The Atlantic flagged it as AI-generated; the author denies using AI","summary":"A week after a Zimbabwean writer won the £10,000 2026 Caine Prize for African Writing, The Atlantic published an article on Sunday 11 October 2026 whose central evidence was a scan by Pangram, an AI-text detection tool, flagging almost the entire winning story as AI-generated. The Caine Prize said it had begun a review: according to its chair, the author would send documentation of his drafts and the prize was talking to the judges and to the magazine that first published the story. The chair told the BBC that The Atlantic had published before a full investigation, tainting the prize and the winner, and said a detector should not be the final arbiter. The author denied using AI and said he had a documented history of the story's development; his literary agency said it was confident the story was his own work. No finding has been announced.","incidentDate":"2026-10-11","incidentKind":"single_event","incidentDatePrecision":"day","exposurePattern":"unknown","reportedDate":"2026-10-13","aiSystem":"Pangram, an AI-text detection tool, used by The Atlantic's journalist to scan the winning story (BBC; Brittle Paper); the alleged use of generative AI to write the story is disputed","aiProduct":"Pangram AI-text detector","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["reputational_harm","professional_harm"],"harmOutcomeSummary":"A prize-winning author was publicly accused, on the strength of an AI-text detector's scan reported by The Atlantic, of submitting an AI-written story; the Caine Prize opened a review of the winning story and asked for documentation of his drafts, and the prize chair said the article had tainted the winner (BBC; Brittle Paper). The author denies using AI.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"One person: the prize winner whose story was flagged by the detector and placed under the prize's review (BBC; Brittle Paper). Writers who commented on social media are not counted.","victimAgeRange":"adult","platformType":"other","outcomeType":"investigation_opened","outcomeStatus":"ongoing","primarySourceUrl":"https://www.bbc.co.uk/news/articles/cm4g1rekpgnjo","primarySourceLabel":"BBC News, 13 October 2026: Zimbabwean author denies using AI to write award-winning story","firstPublishedAt":"2026-10-14T03:12:41.50207+00:00","updatedAt":"2026-10-14T03:12:41.50207+00:00","scopeVersion":"facts-v3","tags":["ai-text-detector","pangram","literary-prize","caine-prize","ai-authorship-accusation","disputed"]},{"id":"2026-karachi-teen-claude-fiction-queries-flagged-fbi-referral-detained-released","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'AI safety guide deemed question of the youth a violation of its safety code and informed the US investigation agency FBI about the suspicious activity.'; 'The US agency informed Pakistani agency about the suspected youth'","relation":"supports","source_id":"s1"}],"assertion":"Officials said the AI's safety system deemed the youth's question a violation of its safety code and informed the FBI, which informed the Pakistani agency.","causal_attribution":"Officials quoted by one outlet; the provider has not described its report."},{"id":"c2","status":"reported","evidence":[{"locator":"'According to the NCCIA, the platform’s safety system flagged the queries on September 4.'; 'The referral prompted the NCCIA to detain the youth for questioning.'","relation":"supports","source_id":"s2"}],"assertion":"According to the NCCIA as reported by ProPakistani, the platform's safety system flagged the queries on 4 September, the information was reportedly shared with the FBI, and the referral prompted the NCCIA to detain the youth for questioning.","causal_attribution":"NCCIA account relayed by one outlet."},{"id":"c3","status":"reported","evidence":[{"locator":"'The News reported that the chatbot flagged the conversation as a breach of its safety safeguards and that Anthropic subsequently reported the activity to the FBI.'","relation":"supports","source_id":"s3"}],"assertion":"The News, as relayed by Bloom Pakistan, reported that the chatbot was Anthropic's Claude, that the chatbot flagged the conversation and that Anthropic reported the activity to the FBI.","causal_attribution":"Relay of one outlet; The News was not read directly."},{"id":"c4","status":"reported","evidence":[{"locator":"'and detained the suspect'; 'The youth was released after completing investigations'","relation":"supports","source_id":"s1"},{"locator":"'Authorities released him after questioning'","relation":"supports","source_id":"s2"},{"locator":"'He was questioned by the NCCIA and later released.'","relation":"supports","source_id":"s3"}],"assertion":"The NCCIA carried out a raid, detained the youth and released him after its investigation.","causal_attribution":"Reported by three outlets that draw on NCCIA officials or investigators; treated as one official origin."},{"id":"c5","status":"reported","evidence":[{"locator":"'He told us that he was working over a fictional crime story'; 'No incident of such type happened with the youth in reality'","relation":"supports","source_id":"s1"}],"assertion":"Officials said the youth told investigators he was working on a fictional crime story, and that nothing of the kind had happened in reality.","causal_attribution":"NCCIA official quoted by one outlet."},{"id":"c6","status":"reported","evidence":[{"locator":"'Investigators ultimately did not establish grounds for prosecution and released the teenager'","relation":"supports","source_id":"s3"}],"assertion":"The News, as relayed by Bloom Pakistan, reported that investigators did not establish grounds for prosecution but remained concerned and were monitoring the situation.","causal_attribution":"Relay of one outlet."}],"effects":[{"label":"raided, detained and questioned after his chatbot queries for a story were flagged and referred via the FBI","claim_id":"c4","direction":"negative"}],"sources":[{"id":"s1","url":"https://arynews.tv/youth-faces-detention-seeking-help-from-ai-fictional-murder","kind":"news_report","access":"read","language":"en","translation_note":"English-language source; no translation. ARY News, 29 September 2026, read live (HTTP 200); quotes NCCIA officials.","independence_group":"nccia-officials-karachi-2026-09"},{"id":"s2","url":"https://propakistani.pk/2026/10/02/karachi-youth-detained-after-claude-queries-trigger-fbi-referral/","kind":"news_report","access":"read","language":"en","translation_note":"English-language source; no translation. ProPakistani, 2 October 2026; text read from the Internet Archive capture (HTTP 200); cites the NCCIA.","independence_group":"nccia-officials-karachi-2026-09"},{"id":"s3","url":"https://bloompakistan.com/karachi-teen-detained-ai-crime-chat-fbi-alert/","kind":"news_report","access":"read","language":"en","translation_note":"English-language source; no translation. Bloom Pakistan, 29 September 2026, read live (HTTP 200); relays The News, which was not read directly.","independence_group":"thenews-karachi-2026-09-29"}],"version":1,"ai_roles":["own_use"],"contexts":["justice","privacy"],"unknowns":["The date of the raid and detention, and how long he was held","What Claude replied to the queries","How the provider reviewed and reported the conversation; Anthropic and the FBI have not commented in the sources read","Whether the NCCIA took or plans any further action"],"geography":{"basis":"All three sources place the user, the raid and the detention in Karachi, Pakistan. The FBI referral passed through the United States, but no event there is described. No court proceeding is reported.","court_countries":[],"event_countries":["PK"],"affected_person_countries":["PK"]},"publication":{"basis":"Published as a reported case: NCCIA officials quoted by ARY News and the NCCIA account reported by ProPakistani describe a chatbot safety flag on a teenager's queries for a crime story, a referral through the FBI, and his detention, questioning and release; Bloom Pakistan, relaying The News, names Claude and Anthropic's report. All claims rest on officials' accounts and are reported.","reviewed_on":"2026-10-14"},"ai_involvement":{"basis":"Officials quoted by ARY News say the AI's safety system deemed the youth's question a violation of its safety code and informed the FBI; ProPakistani, citing the NCCIA, says the platform's safety system flagged the queries on 4 September and the referral prompted the detention. The automated flag is the AI system's judgment about the user (made_claim_about); his exchanges with the chatbot are what it flagged (communicated_with). The News, as relayed by Bloom Pakistan, names the chatbot as Claude and says Anthropic reported the activity to the FBI. What Claude replied and how the provider's report was made are not described.","status":"reported"},"person_relations":["communicated_with","made_claim_about"]},"name":"Karachi, Pakistan: teenager detained and released after Claude flagged his fiction-writing queries and the report reached the FBI, officials say","summary":"A teenager in Karachi was detained by Pakistan's National Cyber Crime Investigation Agency (NCCIA) and later released after his questions to an AI chatbot about how a character in a crime story could be killed were flagged, according to officials quoted by ARY News (29 September 2026) and the NCCIA as cited by ProPakistani (2 October 2026). ProPakistani reports that, according to the NCCIA, the platform's safety system flagged the queries on 4 September and the information was reportedly shared with the US FBI, which passed it to Pakistani authorities; ARY reports that the NCCIA carried out a raid and detained him. Bloom Pakistan, relaying The News, names the chatbot as Anthropic's Claude and reports that Anthropic reported the activity to the FBI. Officials said he told investigators he was writing a fictional crime story, and he was released after questioning; Bloom, relaying The News, says investigators found no grounds for prosecution but remained concerned.","incidentDate":"2026-09-01","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"repeated_interactions","reportedDate":"2026-09-29","aiSystem":"Claude (Anthropic), named by an NCCIA official quoted by ARY News, by ProPakistani and by The News as relayed by Bloom Pakistan; the platform's safety system flagged the teenager's queries, and the activity was reported to the FBI. Who at the provider made the report, and what Claude replied, are not described.","aiProduct":"Claude (reported)","aiCompany":"Anthropic","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["loss_of_liberty","legal_harm"],"harmOutcomeSummary":"According to NCCIA officials quoted by ARY News and ProPakistani, a teenager whose AI-chatbot queries for a crime story were flagged and passed to Pakistani authorities through the FBI was detained in an NCCIA raid and questioned, then released after investigators accepted that the queries concerned a fictional story.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"One teenager detained and released, per ARY News (\"detained the suspect\"), ProPakistani (\"detain the youth for questioning\") and Bloom Pakistan relaying The News. He was the Claude user. No one else is reported to have suffered an adverse consequence.","victimAgeRange":"minor","jurisdiction":"PK","platformType":"chatbot","outcomeType":"investigation_opened","outcomeStatus":"resolved","primarySourceUrl":"https://arynews.tv/youth-faces-detention-seeking-help-from-ai-fictional-murder","primarySourceLabel":"ARY News, 29 September 2026: youth faces detention after seeking help from AI for a fictional murder","firstPublishedAt":"2026-10-14T03:11:18.382947+00:00","updatedAt":"2026-10-14T03:11:18.382947+00:00","scopeVersion":"facts-v3","tags":["claude","anthropic","law-enforcement-report","fbi","detention","released-without-charge","minor","pakistan","fiction-writing"]},{"id":"2026-switzerland-grok-generated-sexist-insults-about-finance-minister-criminal-complaint-bern-investigation","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'Am 10. März hat es auch eine Schweizer Politikerin erwischt: Bundesrätin Karin Keller-Sutter. Es war der Schweizer Peter P.*, der Grok im Verlauf einer generellen «Roast»-Diskussion mit einem Prompt gezielt animierte'","relation":"supports","source_id":"s1"},{"locator":"'Ein Nutzer hatte den Bot am 10. März gezielt zu Beleidigungen gegen Keller-Sutter aufgefordert.'","relation":"supports","source_id":"s2"},{"locator":"'A user of the social network X prompted the AI Grok chatbot to conjure insults against the minister.'","relation":"supports","source_id":"s3"}],"assertion":"On 10 March 2026 a Swiss X user, during a roast thread, prompted Grok to attack Federal Councillor Karin Keller-Sutter in hard street slang.","causal_attribution":"CH Media's report of the prompt from the post it saw; Keystone-SDA relays Sonntagsblick's report of a user prompting Grok. The prompter's identity is a pseudonym in the sources and is not recorded here."},{"id":"c2","status":"reported","evidence":[{"locator":"'Mit vulgären Beschimpfungen. Mit beleidigenden Körper- und Charakterangriffen. Mit abwertender und sexualisierter Sprache.' and 'Die sexualisierten Begriffe von Groks Entgleisungen sind im Screenshot des Posts auf X verpixelt.'","relation":"supports","source_id":"s1"},{"locator":"'Grok reagierte mit sexistischen und beleidigenden Aussagen, die öffentlich einsehbar und teilbar waren'","relation":"supports","source_id":"s2"}],"assertion":"Grok replied on X with vulgar insults, attacks on her body and character, and degrading, sexualised language about Keller-Sutter, and the reply was publicly visible and shareable.","causal_attribution":"CH Media's description of Grok's post from a screenshot; both articles come from the same media group, so the claim rests on one reporting chain."},{"id":"c3","status":"reported","evidence":[{"locator":"'Keller-Sutter erfuhr am Folgetag davon. Kurz darauf wurde der Beitrag gelöscht.' and '«Es war eine harmlose technische Übung, um zu sehen, was mit diesem Grok möglich ist»'","relation":"supports","source_id":"s2"},{"locator":"'Er löschte seinen «Roast»-Prompt und auch die Antwort von Grok.'","relation":"supports","source_id":"s1"}],"assertion":"Keller-Sutter learned of the post the next day, and the user deleted his prompt and Grok's reply shortly afterwards; the user later told Tamedia it was a harmless technical exercise.","causal_attribution":"CH Media reporting; the user's comment is his own account relayed from Tamedia."},{"id":"c4","status":"reported","evidence":[{"locator":"'the Bern public prosecutor's office has opened an investigation over insult and defamation.' and 'The minister’s criminal complaint targets “unknown persons”.'","relation":"supports","source_id":"s3"},{"locator":"'le Ministère public bernois a ouvert une enquête pour injures et diffamation.'","relation":"supports","source_id":"s4"},{"locator":"'Inzwischen hat die Bundesrätin Strafanzeige wegen Verleumdung und Beschimpfung eingereicht.'","relation":"supports","source_id":"s2"}],"assertion":"Keller-Sutter filed a criminal complaint against unknown persons, and the Bern public prosecutor's office confirmed to Keystone-SDA that it had opened an investigation for insult and defamation.","causal_attribution":"The prosecutor's confirmation to Keystone-SDA and statements by Keller-Sutter's spokesperson; the investigation has reached no reported finding."},{"id":"c5","status":"reported","evidence":[{"locator":"'«Solcher Frauenhass darf nicht als normal oder akzeptabel empfunden werden», wurde Hollenstein zitiert.'","relation":"supports","source_id":"s2"},{"locator":"'«Der Verbreitung solcher sexistischer Texte muss eine Grenze gesetzt werden'","relation":"supports","source_id":"s1"}],"assertion":"Keller-Sutter's spokesperson said such misogyny must not be seen as normal or acceptable, and the finance department said the spread of such sexist texts must be limited.","causal_attribution":"Official statements by the minister's spokesperson and department."}],"effects":[{"label":"sexist and degrading insults about a named minister generated and posted publicly by a chatbot","claim_id":"c2","direction":"negative"},{"label":"criminal complaint and prosecutor investigation for insult and defamation","claim_id":"c4","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.watson.ch/schweiz/bundesrat/406055717-ki-chatbot-grok-von-elon-musk-beschimpft-keller-sutter-prozess-droht","kind":"news_report","access":"read","language":"de","translation_note":"Read in German on 2026-10-12 from the publisher page (full article text); the research agent (an AI) translated the passages relied on into English for this record, and no human translator reviewed them. The article credits bzbasel.ch (CH Media).","independence_group":"ch-media"},{"id":"s2","url":"https://www.aargauerzeitung.ch/schweiz/keller-sutter-reicht-strafanzeige-gegen-elon-musks-ki-grok-ein-ld.4147350","kind":"news_report","access":"read","language":"de","translation_note":"Read in German on 2026-10-12 from the publisher page (full article text); the research agent (an AI) translated the passages relied on into English for this record, and no human translator reviewed them. Same newsroom group as s1 (CH Media); it relays Tamedia for the complaint and the user's comment.","independence_group":"ch-media"},{"id":"s3","url":"https://www.swissinfo.ch/eng/various/ai-chatbot-criminal-investigation-following-karin-keller-sutters-complaint/91249480","kind":"news_report","access":"read","language":"en","translation_note":"Read in English on 2026-10-12 from the publisher page; swissinfo.ch states that it translates selected news agency copy with automatic translation tools and that a journalist then reviews the translation, so the English text is a reviewed machine translation of a Keystone-SDA report. No further translation by the research agent (an AI) was needed.","independence_group":"keystone-sda"},{"id":"s4","url":"https://www.swissinfo.ch/fre/chatbot-dia%3aenqu%c3%aate-p%c3%a9nale-apr%c3%a8s-la-plainte-de-karin-keller-sutter/91248423","kind":"news_report","access":"read","language":"fr","translation_note":"Read in French on 2026-10-12 from the publisher page (Keystone-ATS copy); the research agent (an AI) translated the passages relied on into English for this record, and no human translator reviewed them. Same agency as s3.","independence_group":"keystone-sda"}],"version":1,"ai_roles":["others_use"],"contexts":["public_services","everyday_life"],"unknowns":["The full text of Grok's reply as posted; the sources quote parts and describe the rest from a screenshot.","How many people saw or shared the reply before it was deleted.","Whether the investigation has identified anyone or reached any decision, and whether xAI or X responded.","Keller-Sutter's own description of any personal effect; only her spokesperson and department are quoted.","Where the user was when he prompted Grok."],"geography":{"basis":"Keller-Sutter is a member of the Swiss Federal Council and head of its finance department, the office through which the reply targeted her, and the Bern public prosecutor's office opened the investigation. The sources call the user Swiss but do not say where he prompted Grok or where the reply was posted from, so the event country is left unknown; the investigation is not a court proceeding.","court_countries":[],"event_countries":[],"affected_person_countries":["CH"]},"publication":{"basis":"Published on CH Media's description, from a screenshot, of Grok's own reply on X insulting Federal Councillor Keller-Sutter in sexist and degrading terms after a user asked for a roast, together with Keystone-SDA's report that the Bern prosecutor opened an investigation for insult and defamation on her complaint. The generated reply is the AI output and the public abuse of the named minister is the reported harm. The case adds a criminal investigation by a public prosecutor as a legal outcome. Claims stay reported because each rests on one reporting chain or on official statements.","reviewed_on":"2026-10-12"},"ai_involvement":{"basis":"watson (CH Media) describes Grok's own reply on X, from a screenshot of the post: after a user asked for a hard street-slang roast of the minister, Grok answered that it would and wrote vulgar insults, attacks on her body and character, and degrading, sexualised language about her (\"Mit vulgären Beschimpfungen. Mit beleidigenden Körper- und Charakterangriffen. Mit abwertender und sexualisierter Sprache.\"). The Aargauer Zeitung reports that Grok's statements were publicly visible and shareable, and Keystone-SDA reports that a user prompted Grok to produce insults against the minister. The generated reply is the material her criminal complaint for insult and defamation concerns.","status":"reported"},"person_relations":["made_claim_about"]},"name":"Switzerland: Grok posted sexist insults about Finance Minister Karin Keller-Sutter on X at a user's prompt; her complaint led to a criminal investigation","summary":"On 10 March 2026, according to CH Media reporting published by watson, a Swiss X user taking part in a \"roast\" thread asked xAI's chatbot Grok to attack Federal Councillor Karin Keller-Sutter, Switzerland's finance minister, in hard street slang. Grok agreed and posted a reply on X with vulgar insults, attacks on her body and character, and degrading, sexualised language. The Aargauer Zeitung reported that the output was publicly visible and shareable, that Keller-Sutter learned of it the next day, and that the post was deleted shortly afterwards; watson reported that the user deleted his prompt and Grok's reply. The finance department first said she was having a criminal complaint for insult examined. Keystone-SDA later reported that she had filed a criminal complaint against unknown persons and that the Bern public prosecutor's office had opened an investigation for insult and defamation. Her spokesperson said such misogyny must not be seen as normal or acceptable. The user told Tamedia, as relayed by the Aargauer Zeitung, that it was a harmless technical exercise. No outcome of the investigation is reported.","incidentDate":"2026-03-10","incidentKind":"single_event","incidentDatePrecision":"day","exposurePattern":"single_interaction","reportedDate":"2026-03-15","aiSystem":"Grok, the xAI chatbot that replies to users on X, prompted by an X user during a \"roast\" thread","aiProduct":"Grok","aiCompany":"xAI","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["reputational_harm"],"harmOutcomeSummary":"Grok's reply, posted publicly on X, subjected Keller-Sutter to sexist and degrading insults that the FDP co-president who discussed it with her called a sexist insult of the ugliest and most vulgar kind. Keller-Sutter filed a criminal complaint for insult and defamation, and her spokesperson described the content as misogyny. Her own description of any personal effect is not reported.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"One person: Karin Keller-Sutter, the target of Grok's reply, who filed the complaint. Other politicians targeted in other users' roast threads are not part of this event and are not counted.","victimAgeRange":"adult","jurisdiction":"CH","platformType":"chatbot","outcomeType":"investigation_opened","outcomeStatus":"ongoing","primarySourceUrl":"https://www.watson.ch/schweiz/bundesrat/406055717-ki-chatbot-grok-von-elon-musk-beschimpft-keller-sutter-prozess-droht","primarySourceLabel":"watson, 15 March 2026: KI-Chatbot Grok von Elon Musk beschimpft Keller-Sutter, Prozess droht","firstPublishedAt":"2026-10-12T03:13:18.439789+00:00","updatedAt":"2026-10-12T03:13:18.439789+00:00","scopeVersion":"facts-v3","tags":["grok","x","chatbot","insult","defamation","misogyny","switzerland","criminal-complaint","public-official","made-claim-about"]},{"id":"2024-tasmania-parole-board-ai-hallucinated-case-law-media-condition-quashed","caseFacts":{"claims":[{"id":"c1","status":"corroborated","evidence":[{"locator":"In December 2024, the Parole Board of Tasmania added a condition that she not speak \"directly or indirectly with any media outlet to claim (her) alleged innocence and/or wrongful conviction\".","relation":"supports","source_id":"s1"},{"locator":"This was replaced by a new parole condition in May 2025","relation":"supports","source_id":"s1"},{"locator":"The condition was tweaked in 2025 to prevent her from communicating directly or indirectly - including through third parties and written statements - with media about her alleged innocence","relation":"supports","source_id":"s2"},{"locator":"The Supreme Court declared that conditions imposed by the Parole Board of Tasmania in December 2024 and May 2025 were invalid and unlawful.","relation":"supports","source_id":"s3"}],"assertion":"In December 2024 the Parole Board of Tasmania added a parole condition barring the parolee from communicating with media to claim her innocence or wrongful conviction, and a broader condition replaced it in May 2025.","causal_attribution":"Not a causal claim."},{"id":"c2","status":"reported","evidence":[{"locator":"used artificial intelligence to prepare a rationale justifying the parole condition, citing case law that was \"fictitious\" and likely \"AI hallucinations\".","relation":"supports","source_id":"s1"},{"locator":"relied upon by the board in creating the conditions, included erroneous AI information, the court was told.","relation":"supports","source_id":"s2"},{"locator":"\"As we now know, those cases didn't exist,\" he said.","relation":"supports","source_id":"s2"}],"assertion":"Her counsel told the Supreme Court that the board's secretary used artificial intelligence to prepare a rationale document justifying the condition and that the case law it cited did not exist.","causal_attribution":"Counsel's statement in court is reported by ABC and AAP. These accounts do not independently verify the generation history or identify the tool. The document itself was not inspected."},{"id":"c3","status":"reported","evidence":[{"locator":"the Board’s decision-making was affected by a document containing legal authorities that were likely AI-generated hallucinations and did not in fact exist","relation":"supports","source_id":"s3"},{"locator":"Justice Estcourt also found that when the parole board made the decision to impose this condition, it had relied on a document that cited case law that was \"fictitious\" and likely \"AI hallucinations\".","relation":"supports","source_id":"s4"}],"assertion":"The Human Rights Law Centre says the State conceded the board's decision-making was affected by a document containing nonexistent authorities likely generated by AI. ABC's 19 September report says the court found the board had relied on the document when imposing the condition.","causal_attribution":"The State's concession as reported by the applicant's representative, and the ABC's report of the court's finding. The ABC's 15 September report says the judge heard this; its 19 September report says the judge found it. The AI origin is stated as likely."},{"id":"c4","status":"corroborated","evidence":[{"locator":"Today in the Supreme Court in Hobart, Justice Stephen Estcourt ruled the condition was invalid because it denied her procedural fairness, was legally unreasonable, and created uncertainty.","relation":"supports","source_id":"s1"},{"locator":"Justice Stephen Estcourt found the condition was made without procedural fairness, was legally unreasonable and was filled with uncertainty.","relation":"supports","source_id":"s2"}],"assertion":"ABC and AAP report that on 15 September 2026 Justice Stephen Estcourt of the Supreme Court of Tasmania ruled the condition invalid because it denied procedural fairness, was legally unreasonable and created uncertainty.","causal_attribution":"The reported grounds of invalidity are procedural fairness, unreasonableness and uncertainty; the AI document is not reported as a separate ground."},{"id":"c5","status":"corroborated","evidence":[{"locator":"The parole board decided to withdraw the condition in August and did not contest it being set aside on Tuesday.","relation":"supports","source_id":"s2"},{"locator":"Mr Coleridge said the parole board met and withdrew the condition that same month","relation":"supports","source_id":"s1"}],"assertion":"The board withdrew the condition in August 2026 and did not contest it being set aside.","causal_attribution":"Not a causal claim."},{"id":"c6","status":"reported","evidence":[{"locator":"The court heard that in August","relation":"supports","source_id":"s1"},{"locator":"brought to the court's attention that the case law referred to in the document did not exist.","relation":"supports","source_id":"s1"}],"assertion":"Her lawyers raised in August 2026 that the case law in the document did not exist, and the board withdrew the condition that month.","causal_attribution":"Counsel's sequence of events; the board refused to explain the withdrawal."},{"id":"c7","status":"reported","evidence":[{"locator":"For almost two years, the conditions imposed by the Parole Board of Tasmania prevented Ms Neill-Fraser from speaking publicly about her claimed innocence","relation":"supports","source_id":"s3"},{"locator":"she had been unable to speak publicly after Tasmania's parole board placed a condition on her release in 2024.","relation":"supports","source_id":"s2"},{"locator":"Any breach of parole could result in the remainder of the sentence having to be served","relation":"supports","source_id":"s6"}],"assertion":"The Human Rights Law Centre says the conditions prevented the parolee from speaking publicly about her claimed innocence for almost two years. The Guardian reports that a parole breach could have meant serving the remaining sentence.","causal_attribution":"The restriction is the effect of the condition; the AI document is one input to the decision that imposed it."},{"id":"c8","status":"reported","evidence":[{"locator":"\"The Parole Board's professional standards will be independently investigated following its use of unverified artificial intelligence in a rationale document which formed part of the materials before the board to impose conditions,\"","relation":"supports","source_id":"s4"}],"assertion":"Tasmania's Attorney-General announced an independent investigation of the Parole Board's professional standards following its use of unverified artificial intelligence in the rationale document.","causal_attribution":"The Attorney-General's statement acknowledges the AI use; it does not quantify its effect on the decision."},{"id":"c9","status":"corroborated","evidence":[{"locator":"\"The Department will undertake a review to determine the extent to which Artificial Intelligence (AI) may have been used to inform past Tasmanian Parole Board decisions,\"","relation":"supports","source_id":"s4"},{"locator":"“The department will undertake a review to determine the extent to which artificial intelligence may have been used to inform past Tasmanian Parole Board decisions,” a justice department spokesperson said.","relation":"supports","source_id":"s5"}],"assertion":"Tasmania's Justice Department said it would review the extent to which AI may have been used to inform past Parole Board decisions.","causal_attribution":"Not a causal claim."}],"effects":[{"label":"a restriction on public speech about claimed innocence from December 2024 until withdrawal in August 2026","claim_id":"c7","direction":"negative"},{"label":"adverse parole decision affected by a rationale document citing likely AI-generated, non-existent case law","claim_id":"c3","direction":"negative"}],"sources":[{"id":"s1","url":"https://www.abc.net.au/news/2026-09-15/susan-neill-fraser-media-ban-parole-board-ai-hallucinations/106302950","kind":"news_report","access":"read","language":"en","translation_note":"Read in English in full from abc.net.au (HTTP 200). Court reporting of the 15 September 2026 hearing; the judgment and orders were not inspected. Retained complete body independently reinspected on 2026-10-06 for this proposal. No translation required.","independence_group":"abc-au"},{"id":"s2","url":"https://www.canberratimes.com.au/story/9350990/murderers-gag-order-lifted-after-troubling-ai-error/","kind":"news_report","access":"read","language":"en","translation_note":"Read in English in full from canberratimes.com.au (HTTP 200); Australian Associated Press copy of the same hearing, reported independently of the ABC. Retained complete body independently reinspected on 2026-10-06 for this proposal. No translation required.","independence_group":"aap"},{"id":"s3","url":"https://www.hrlc.org.au/news/susan-neill-fraser-gag-order-unlawful/","kind":"organisation_statement","access":"read","language":"en","translation_note":"Read in English in full from hrlc.org.au (HTTP 200). Release by the organisation that represented the parolee; its account of the State's concessions is a party-side account. Retained complete body independently reinspected on 2026-10-06 for this proposal. No translation required.","independence_group":"hrlc"},{"id":"s4","url":"https://www.abc.net.au/news/2026-09-19/parole-board-ai-use-review-after-neill-fraser-case/107172064","kind":"news_report","access":"read","language":"en","translation_note":"Read in English in full from abc.net.au (HTTP 200). Same outlet as s1, so the same independence group; carries the Attorney-General's and Justice Department's statements. Retained complete body independently reinspected on 2026-10-06 for this proposal. No translation required.","independence_group":"abc-au"},{"id":"s5","url":"https://www.theguardian.com/australia-news/2026/sep/19/tasmania-justice-department-review-ai-artificial-intelligence-fake-citation-murderer-parole-decision-ntwnfb","kind":"news_report","access":"read","language":"en","translation_note":"Read in English in full from theguardian.com (HTTP 200); credited to a staff reporter with Australian Associated Press, so grouped with s2. Retained complete body independently reinspected on 2026-10-06 for this proposal. No translation required.","independence_group":"aap"},{"id":"s6","url":"https://www.theguardian.com/australia-news/2026/sep/26/sue-neill-fraser-yacht-murder-ai-court-case-media-ban-tasmania-ntwnfb","kind":"news_report","access":"read","language":"en","translation_note":"Read in English in full from theguardian.com (HTTP 200); Guardian Australia feature by its own reporter. Retained complete body independently reinspected on 2026-10-06 for this proposal. No translation required.","independence_group":"guardian-au"}],"version":1,"ai_roles":["institutional_use"],"contexts":["justice"],"unknowns":["Which AI tool was used and the extent of its contribution to the rationale. Counsel identified the secretary as its preparer, but the document and generation history were not inspected.","Whether the rationale document informed the December 2024 decision, the May 2025 decision or both.","The text of the judgment or orders, which could not be read (AustLII HTTP 403).","The findings of the independent investigation and of the Justice Department's review of past parole decisions."],"geography":{"basis":"The Parole Board of Tasmania imposed the condition and the Supreme Court of Tasmania in Hobart set it aside (ABC, AAP). The parolee's country of residence is not stated as such in the sources read.","court_countries":["AU"],"event_countries":["AU"],"affected_person_countries":[]},"publication":{"basis":"This is a concrete restriction of a person's speech through a human parole-board decision. Court reporting and the representative's account describe likely AI-generated false authorities used as an input to that decision. The relation is informed_decision_about and does not require the AI text to mention the parolee. The AI origin remains qualified, and the document's precise role in each of the two successive conditions is unknown. The court's reported grounds concern procedural fairness, unreasonableness and uncertainty.","reviewed_on":"2026-10-06"},"ai_involvement":{"basis":"The Human Rights Law Centre reports the State's concession that the board's decision-making was affected by a rationale document containing nonexistent authorities likely generated by AI. ABC and AAP court reports describe the document's use in imposing the media condition, and the Attorney-General's statement reported by ABC acknowledges unverified AI in material used by the board. These sources connect the AI-supplied legal material to a human decision restricting the parolee's speech. The relation is informed_decision_about. The sources do not establish that AI made claims about the parolee herself, wrote the whole rationale or made the final parole decision. The judgment and the rationale document have not been inspected.","status":"supported"},"person_relations":["informed_decision_about"]},"name":"Tasmania parole media ban quashed after state conceded reliance on likely AI-hallucinated case law","summary":"In December 2024 the Parole Board of Tasmania added a condition to the parole of Susan Neill-Fraser, released in 2022 after serving 13 years for murder and who maintains her innocence, barring her from communicating with media to claim innocence or wrongful conviction; a broader condition replaced it in May 2025. On 15 September 2026 the Supreme Court of Tasmania declared the conditions invalid. Her counsel told the court that the board's secretary had used artificial intelligence to prepare a rationale document justifying the condition and that it cited case law that did not exist. The Human Rights Law Centre, which represented her, says the State conceded that the board's decision-making was affected by a document containing legal authorities that were likely AI-generated hallucinations, and the ABC reports the court found the board relied on that document. The board had withdrawn the condition in August 2026. Tasmania's Attorney-General announced an independent investigation of the board's professional standards over its use of unverified AI, and the Justice Department announced a review of AI use in past parole decisions.","incidentDate":"2024-12-01","incidentEndDate":"2026-09-15","incidentKind":"bounded_series","incidentDatePrecision":"month","exposurePattern":"unknown","reportedDate":"2026-09-15","aiSystem":"An unidentified artificial-intelligence tool that, according to counsel and the Attorney-General, was used to prepare a Parole Board rationale document citing case law that did not exist","aiProduct":"Unidentified AI writing tool","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["legal_harm","loss_of_autonomy"],"harmOutcomeSummary":"The conditions restricted a parolee's public speech from December 2024 until their withdrawal in August 2026. They were declared invalid in September. The Human Rights Law Centre says the State conceded the board's decision-making had been affected by nonexistent legal authorities likely generated by AI. The judgment was not directly inspected, and the sources do not establish whether that document informed the original condition, the replacement or both.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"One parolee is counted for the reported restriction of her speech. The reviewed sources describe no other identified person harmed by AI use in a parole decision; an announced review of past decisions does not establish additional victims.","victimAgeRange":"elderly","jurisdiction":"AU","platformType":"other","outcomeType":"investigation_opened","outcomeStatus":"ongoing","primarySourceUrl":"https://www.abc.net.au/news/2026-09-15/susan-neill-fraser-media-ban-parole-board-ai-hallucinations/106302950","primarySourceLabel":"ABC News, 15 September 2026: AI hallucinations in parole board submission scuttle media ban on Susan Neill-Fraser","firstPublishedAt":"2026-10-06T04:48:51.9027+00:00","updatedAt":"2026-10-06T04:48:51.9027+00:00","scopeVersion":"facts-v3","tags":["parole","hallucinated-case-law","freedom-of-speech","judicial-review","government-ai-use"]},{"id":"2025-stephens-county-oklahoma-judge-chatgpt-fake-citations-paternity-order","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'Wheeler denied a request from one of the parents' attorneys to have the other parent undergo a psychological test—and in his order denying the request, he pointed to multiple previous Oklahoma cases he said backed him up.'","relation":"supports","source_id":"s2"},{"locator":"'The 2025 order in a child paternity case'; 'He came under scrutiny after he reprimanded attorney [the attorney] last November in an order in the paternity case for asking for a psychological evaluation of the child’s father.'","relation":"supports","source_id":"s1"},{"locator":"'The issue came to light after a request for a mental evaluation was denied and a subsequent court order was filed.'","relation":"supports","source_id":"s3"}],"assertion":"In November 2025, in a child paternity and custody case, Judge Wheeler issued an order denying a request by one parent's attorney for a psychological evaluation of the other parent, citing Oklahoma cases he said supported the decision.","causal_attribution":"Court history as reported by The Oklahoman and KFOR and described by the mother; the order itself was not read."},{"id":"c2","status":"reported","evidence":[{"locator":"'The judge told the OSBI he only used ChatGPT for research and created the order himself, DA Jason Hicks wrote. The judge, however, also “indicated that at least two of the case citations included in his order and produced by ChatGPT do not exist.”'","relation":"supports","source_id":"s1"},{"locator":"'Stephens County District Attorney Jason Hicks says Wheeler admitted to OSBI agents he cited \"at least two\" cases ChatGPT fed him that \"do not exist.\"'","relation":"supports","source_id":"s2"},{"locator":"'Hicks wrote that Wheeler acknowledged using case citations generated by ChatGPT and that at least two citations included in the order did not exist.'","relation":"supports","source_id":"s3"},{"locator":"'Associate District Judge Lawrence Wheeler used OpenAI's ChatGPT for research while drafting the order, which cited two cases that do not exist, according to state records.'","relation":"supports","source_id":"s4"}],"assertion":"According to the district attorney's 17 August 2026 letter, the judge told the OSBI that he used ChatGPT for research and wrote the order himself, and indicated that at least two case citations in the order, produced by ChatGPT, do not exist.","causal_attribution":"The judge's own statement to investigators as relayed in the district attorney's letter; all four reports rely on that letter, which was not read. The judge's attorney declined to comment."},{"id":"c3","status":"reported","evidence":[{"locator":"'The judge issued the reprimand “for stooping to such frivolous trial tactics” and cautioned that future similar filings may be subject to further sanctions.'; '[The attorney] challenged the order at the Oklahoma Supreme Court in February. He told justices that Wheeler based his order on two cases, Cummings v. Cimarron Elevator Co. and Hawkins v. Linhart, that do not exist.'; '[The attorney] dismissed his challenge in March after Wheeler vacated the order. Wheeler is no longer involved in the paternity case.'","relation":"supports","source_id":"s1"},{"locator":"'She said her attorney at the time challenged the ruling and began reviewing the legal authorities cited in the order.'","relation":"supports","source_id":"s3"}],"assertion":"The order reprimanded the attorney who had asked for a psychological evaluation of the child's father 'for stooping to such frivolous trial tactics' and warned of further sanctions; that attorney challenged it at the Oklahoma Supreme Court in February 2026, saying it relied on two nonexistent cases, and dismissed the challenge in March after Wheeler vacated the order. The mother told News 9 that her attorney at the time challenged the ruling.","causal_attribution":"The Oklahoman's account of the court filings and the mother's account to News 9; the Supreme Court filings were not read."},{"id":"c4","status":"reported","evidence":[{"locator":"'[The mother] said learning the citations were allegedly fabricated was alarming because the order involved the custody of a child.'; '“Then it became very concerning to me that somebody would do something like that over a child custody case,” she said.'; '“He then turned around and decided to recuse himself from my case,” she said. “He called my attorney that I had at the time and apologized to him.”'","relation":"supports","source_id":"s3"}],"assertion":"The mother told News 9 that learning the citations were allegedly fabricated was alarming because the order involved the custody of a child, and that the judge later recused himself from her case and apologised to her attorney.","causal_attribution":"The mother's first-person account in a broadcast interview."},{"id":"c5","status":"reported","evidence":[{"locator":"'The OSBI investigated the judge at the request of the Oklahoma Council on Judicial Complaints.'; 'The OSBI also looked into allegations “that Judge Wheeler offered to amend the order in exchange for a dismissal of a writ seeking relief from the Oklahoma Supreme Court” and that he “eavesdropped on jury deliberations,” the DA disclosed in the letter.'; 'Prosecutors in the AG’s office “thoroughly reviewed the results of the investigation and determined the evidence does not support a criminal prosecution,” the AG’s press secretary, Leslie Berger, said on Sept. 2.'","relation":"supports","source_id":"s1"},{"locator":"'The Criminal Justice Division of the Attorney General’s Office thoroughly reviewed the results of the investigation and determined the evidence does not support a criminal prosecution. The Oklahoma Supreme Court and the Court on the Judiciary possess ultimate jurisdiction over disciplinary actions for allegations of inappropriate behavior by judges regardless of whether the behavior constitutes a violation of a criminal statute.'","relation":"supports","source_id":"s3"}],"assertion":"The OSBI investigated the judge at the request of the Council on Judicial Complaints, including allegations that he offered to amend the order in exchange for dismissal of the writ and that he eavesdropped on jury deliberations; the attorney general's office said the evidence did not support criminal prosecution and that judicial discipline rests with the Oklahoma Supreme Court and the Court on the Judiciary.","causal_attribution":"District attorney's letter and attorney general's statement as reported; no disciplinary finding has been made."}],"effects":[{"label":"the mother faced an order denying a psychological-evaluation request, with the judge citing authorities reported to have been generated by ChatGPT","claim_id":"c1","direction":"negative"},{"label":"the attorney who requested the evaluation was reprimanded in the order and sought relief from the Oklahoma Supreme Court before the order was vacated","claim_id":"c3","direction":"negative"},{"label":"the mother describes learning of the fabricated citations in her child custody case as alarming","claim_id":"c4","direction":"negative"}],"sources":[{"id":"s1","url":"https://minnlawyer.com/2026/09/09/oklahoma-judge-ai-fabricated-case-citations/","kind":"news_report","access":"read","language":"en","translation_note":"The Oklahoman's report (USA TODAY Network via Reuters Connect) as republished by Minnesota Lawyer; read in full from the Wayback capture of 10 September 2026 because the live page returned 403. Its ChatGPT-admission passages relay the district attorney's letter; its account of the reprimand, the Supreme Court challenge and the vacatur draws on court filings and the attorney. Reinspected in the retained complete English body on 2026-10-06. All sources share the district-attorney-letter chain for AI involvement, so grouped conservatively together; the News 9 maternal interview is separate testimony about her experience, not a second verification of AI use.","independence_group":"wheeler-da-letter"},{"id":"s2","url":"https://www.aol.com/articles/oklahoma-judge-admitted-citing-fake-030000000.html","kind":"news_report","access":"read","language":"en","translation_note":"KFOR (Oklahoma City) report syndicated on AOL, read in full on 2026-09-29. Its account of the ChatGPT admission relays the same district attorney letter. Reinspected in the retained complete English body on 2026-10-06. All sources share the district-attorney-letter chain for AI involvement, so grouped conservatively together; the News 9 maternal interview is separate testimony about her experience, not a second verification of AI use.","independence_group":"wheeler-da-letter"},{"id":"s3","url":"https://www.news9.com/oklahoma-city-news/stephens-county-mother-speaks-out-after-judge-accused-of-using-chatgpt-in-custody-case","kind":"news_report","access":"read","language":"en","translation_note":"News 9 (Griffin Media) interview with the mother, read in full on 2026-09-29. Its description of the ChatGPT admission also relays the district attorney's letter; the mother's own account is independent of it. Reinspected in the retained complete English body on 2026-10-06. All sources share the district-attorney-letter chain for AI involvement, so grouped conservatively together; the News 9 maternal interview is separate testimony about her experience, not a second verification of AI use.","independence_group":"wheeler-da-letter"},{"id":"s4","url":"https://hoodline.com/2026/09/stephens-county-judge-faces-discipline-over-chatgpt-written-court-order/","kind":"news_report","access":"read","language":"en","translation_note":"Hoodline rewrite of the Reuters report, read in full on 2026-09-29. Reuters relays the same district attorney letter. Reinspected in the retained complete English body on 2026-10-06. All sources share the district-attorney-letter chain for AI involvement, so grouped conservatively together; the News 9 maternal interview is separate testimony about her experience, not a second verification of AI use.","independence_group":"wheeler-da-letter"}],"version":1,"ai_roles":["institutional_use"],"contexts":["justice","relationships"],"unknowns":["The text of the order and whether the nonexistent citations were decisive for denying the evaluation and for the reprimand.","Whether any disciplinary proceeding follows from the Council on Judicial Complaints referral.","The current state of the custody case after the vacatur and the judge's departure from it.","The exact day in November 2025 on which the order was issued."],"geography":{"basis":"The Oklahoman, KFOR and News 9 locate the order and proceedings in Stephens County, Oklahoma, and the Oklahoma Supreme Court. These establish the event and court countries as US. News 9 calls the interviewee a Stephens County mother but gives no explicit residence evidence for either counted person, so affected_person_countries is unknown.","court_countries":["US"],"event_countries":["US"],"affected_person_countries":[]},"publication":{"basis":"This reported event connects specific ChatGPT-generated citations to an adverse order issued by a human judge. The Oklahoman and KFOR describe their use in the order, The Oklahoman reports the attorney's reprimand and successful challenge, and News 9 carries the mother's account. The AI admission is reported through one district attorney's letter, not independently verified model output. The family and attorney remain unnamed. No broader custody consequence or disciplinary finding is inferred.","reviewed_on":"2026-10-06"},"ai_involvement":{"basis":"The Oklahoman, KFOR and News 9 relay a district attorney's letter saying the judge acknowledged using ChatGPT for research and incorporating at least two nonexistent citations it produced into his order. The cited authorities were presented in support of the human judge's denial of a requested evaluation and his reprimand of the attorney. This supports informed_decision_about for the affected mother and attorney. It does not establish that ChatGPT itself decided custody or communicated with either of them. The underlying letter and order were not inspected, and whether the citations determined the outcome is unknown.","status":"supported"},"person_relations":["informed_decision_about"]},"name":"Prosecutor: judge used ChatGPT citations in Oklahoma custody order later vacated","summary":"In November 2025 Stephens County Associate District Judge Lawrence Wheeler issued an order in a child paternity and custody case that denied one parent's request for a psychological evaluation of the other parent and reprimanded the requesting parent's attorney 'for stooping to such frivolous trial tactics'. The attorney challenged the order at the Oklahoma Supreme Court in February 2026, telling the justices that it relied on two cases that do not exist; the challenge was dismissed in March after Wheeler vacated the order, and he is no longer on the case. According to a 17 August 2026 letter from the Stephens County district attorney, Wheeler told the Oklahoma State Bureau of Investigation that he used ChatGPT for research and wrote the order himself, and that at least two citations in it produced by ChatGPT do not exist. The mother in the case told News 9 that learning the order's citations were allegedly fabricated was alarming because it concerned the custody of a child. The attorney general's office declined criminal prosecution; judicial discipline remains possible.","incidentDate":"2025-11-01","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"unknown","reportedDate":"2026-09-09","aiSystem":"ChatGPT (OpenAI), used by the judge for legal research while preparing the order","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["legal_harm","professional_harm"],"harmOutcomeSummary":"According to the district attorney's letter as reported by The Oklahoman and KFOR, a judge acknowledged that his custody-case order contained at least two nonexistent citations supplied by ChatGPT. The order denied a request for a psychological evaluation and reprimanded the requesting attorney. The attorney challenged it at the Oklahoma Supreme Court before it was vacated. The mother described the discovery as alarming in a News 9 interview. The order and letter have not been inspected directly.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":2,"affectedCountStatus":"documented_minimum","affectedCountEvidence":"Two people affected through the judge's use of ChatGPT: the mother, whose request for a psychological evaluation was denied by the order and who describes the discovery as alarming (News 9), and the attorney who requested the evaluation and was reprimanded in the order and challenged it at the Oklahoma Supreme Court (The Oklahoman). The child and the other parent are not counted because no harm to them is reported. The judge, the AI user, is not counted: he faces a possible disciplinary process but no sanction has been reported and the complaint bundled other allegations.","victimAgeRange":"adult","jurisdiction":"US-OK","platformType":"assistant","outcomeType":"investigation_opened","outcomeStatus":"ongoing","primarySourceUrl":"https://minnlawyer.com/2026/09/09/oklahoma-judge-ai-fabricated-case-citations/","primarySourceLabel":"The Oklahoman via Minnesota Lawyer, 9 September 2026: judge used ChatGPT citations in custody order","firstPublishedAt":"2026-09-29T09:10:18.787549+00:00","updatedAt":"2026-10-06T04:48:44.069668+00:00","scopeVersion":"facts-v3","tags":["justice","family-court","judge","fabricated-citations","chatgpt","institutional-use","oklahoma"]},{"id":"2026-delhi-police-facial-recognition-jantar-mantar-list-25-in-jail","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'According to the police affidavit submitted on August 17, its FRS identified individuals with criminal records at the protest site'; 'It said that 2,402 people were identified by the Delhi Police'; 'and 471 through criminal records.'; 'No actions are taken solely based upon the result of the facial recognition system which is only the first step to identify a person in the crowd who has a criminal record.'","relation":"supports","source_id":"s1"},{"locator":"'According to the Delhi Police affidavit, 2,402 people were identified through its biometric database, Crime Kundli, while another 471 were identified through criminal records.'","relation":"supports","source_id":"s2"},{"locator":"'The Delhi Police's affidavit of August 17, 2026 told the Supreme Court that 2,873 persons with criminal records had been spotted by the FRS at the protest site.'","relation":"supports","source_id":"s3"}],"assertion":"In an affidavit of 17 August 2026 before the Supreme Court, Delhi Police said its Facial Recognition System had identified 2,873 people with criminal antecedents at the Jantar Mantar protest site between 20 and 26 July 2026, 2,402 through its 'Crime Kundli' biometric database and 471 through criminal records, and that no action is taken solely on an FRS result before field verification.","causal_attribution":"The affidavit as quoted by The Indian Express and relayed by Outlook and IFF; not inspected directly."},{"id":"c2","status":"reported","evidence":[{"locator":"'An investigation of those with serious charges, murder, attempted murder and rape, and child abuse, shows that at least 25 people on the list, identified by the Facial Recognition System (FRS) at Jantar Mantar, were in jail as per police, jail and court records.'; 'In this group, 17 faced charges of murder; four of rape, including two in POCSO cases; and four more of attempt to murder were lodged in Delhi'; 'Court and jail records for all 25 show custody dates precede the protest'","relation":"supports","source_id":"s1"},{"locator":"'At least 25 of the 2,873 people identified by Delhi Police'; 'were already in jail when they were flagged, an investigation of police, prison and court records by The Indian Express has found.'","relation":"supports","source_id":"s2"},{"locator":"'Three were \"identified\" on July 24, twenty-one on July 25 and one on July 26, each with a timestamp.'","relation":"supports","source_id":"s3"}],"assertion":"At least 25 of the 205 listed people facing murder, rape, POCSO or attempted-murder charges were in Tihar, Mandoli or Rohini jails when the FRS logged them at Jantar Mantar (17 murder, four rape of which the Express's narrative counts two and its itemised list three as POCSO cases, four attempted murder), with timestamps on 24, 25 and 26 July 2026 and custody dates that in some cases precede the protest by years.","causal_attribution":"The Indian Express's own check of police, prison and court records (one chain, relayed by Outlook and IFF); the Delhi Police response did not address the 25 cases."},{"id":"c3","status":"reported","evidence":[{"locator":"'On Tuesday, when the Supreme Court quashed all FIRs related to the exam leak protests by students, it let the government proceed against the 2,873 people who the Delhi Police claimed had criminal records'","relation":"supports","source_id":"s1"},{"locator":"'Delhi Police was given permission to register one fresh FIR concerning 2,873 people alleged to have serious criminal antecedents and initially reported to have been present at Jantar Mantar.'; 'The Court did not find that the 2,873 were present, had qualifying criminal records or committed violence.'; 'the order prescribed no method or express requirement that checks precede arrest or other action'","relation":"supports","source_id":"s4"},{"locator":"'On September 1, 2026, a Bench headed by the Chief Justice of India, invoking Article 142, quashed all FIRs registered against students who participated in the July protests'; 'while directing that this should not prejudice their rights and that they must be given every opportunity to defend themselves'","relation":"supports","source_id":"s3"}],"assertion":"On 1 September 2026 the Supreme Court quashed the FIRs against the student protesters but permitted Delhi Police to register a fresh FIR concerning the 2,873 persons with alleged serious antecedents, directing that this should not prejudice their rights, and without a finding that they were present; the undertaking that action would follow field verification is the police's own, and The Federal notes the order prescribed no such requirement.","causal_attribution":"The order as described by the three outlets; not inspected directly."},{"id":"c4","status":"reported","evidence":[{"locator":"'Responding on Monday to a questionnaire from the Express, the Delhi Police said:'; 'During the protests, 2873 individuals identified as prima facie having criminal antecedents were spotted. However, further verification in this regard is pending.'","relation":"supports","source_id":"s1"},{"locator":"'In its response to the Express, the Delhi Police did not address the 25 cases'","relation":"supports","source_id":"s3"}],"assertion":"Responding to The Indian Express, Delhi Police said the 2,873 individuals had been spotted as prima facie having criminal antecedents and that further verification was pending; the response did not address the 25 cases.","causal_attribution":"Police statement as quoted by the Express."},{"id":"c5","status":"reported","evidence":[{"locator":"'In response to a Right to Information (RTI) request in 2022, the police had revealed that it considered a match to be'; 'if its system showed an accuracy rate of 80 per cent.'","relation":"supports","source_id":"s1"},{"locator":"'the Delhi Police disclosed that it treats any match with a similarity score above 80% as \"positive\"'; 'extracted in 2022 only after a Second Appeal before the Central Information Commission'","relation":"supports","source_id":"s3"}],"assertion":"Delhi Police disclosed in 2022, in reply to Right to Information requests, that it treats a facial-recognition match with an 80 per cent similarity or accuracy score as positive.","causal_attribution":"Delhi Police's 2022 RTI replies as described by IFF, which obtained them, and by The Indian Express, which IFF says relies on the same disclosures; one chain. The reply itself was not inspected. IE describes the threshold as an accuracy rate of 80 per cent and IFF as a similarity score above 80 per cent."}],"effects":[{"label":"at least 25 people in judicial custody were logged by the police facial-recognition system as present at a protest and listed in a sworn affidavit of persons with criminal antecedents against whom a fresh FIR and field verification were permitted","claim_id":"c2","direction":"negative"}],"sources":[{"id":"s1","url":"https://indianexpress.com/article/express-exclusive/delhi-police-facial-recognition-jantar-mantar-protest-accused-in-jail-10862099/","kind":"news_report","access":"read","language":"en","translation_note":"Read on 2026-09-24 from an Internet Archive capture of 15 September 2026 (web.archive.org/web/20260915075622id_/...); the live page returns 403 to this host. Express Investigation of 4 September 2026 with the list of the 25 and the police response.","independence_group":"indian-express-investigation"},{"id":"s2","url":"https://www.outlookindia.com/national/jantar-mantar-protest-25-people-flagged-by-facial-recognition-were-in-jail-records-show","kind":"news_report","access":"read","language":"en","translation_note":"Read live on 2026-09-24 (Outlook India, 4 September 2026): relays the Express findings and the affidavit figures; same chain.","independence_group":"indian-express-investigation"},{"id":"s3","url":"https://internetfreedom.in/delhi-polices-facial-recognition-flagged-25-jailed-people-at-jantar-mantar-this-dragnet-must-end/","kind":"other","access":"read","language":"en","translation_note":"Read live on 2026-09-24 (Internet Freedom Foundation, Apar Gupta, 4 September 2026): advocacy analysis that relays the Express findings (same chain for the 25) and adds IFF's own 2022 RTI replies on the 80 per cent threshold, its 24 July 2026 representation and RTI applications.","independence_group":"iff-own-rti"},{"id":"s4","url":"https://thefederal.com/the-federal-special/one-protest-two-probes-why-the-2873-person-fir-raises-questions-255688","kind":"news_report","access":"read","language":"en","translation_note":"Read live on 2026-09-24 (The Federal, 4 September 2026): analysis of the Supreme Court order and the affidavit relying on the Express report; same chain.","independence_group":"indian-express-investigation"}],"version":1,"ai_roles":["institutional_use"],"contexts":["justice","public_services","privacy"],"unknowns":["Whether any of the 25 has been named in the fresh FIR, visited for field verification or otherwise acted against.","Whether the 25 (or the other 2,848 listed people) have been told they are on the list.","How many of the remaining 2,668 entries would fail the same check; the Express examined only the 205 most serious.","The vendor, algorithm and threshold actually applied at Jantar Mantar in July 2026.","The full text of the 17 August affidavit and the 1 September order (described as quoted).","Whether two or three of the four rape accused are POCSO cases: The Indian Express's narrative says two, its itemised list says three.","The protest window: The Indian Express and Outlook give 20 to 26 July; The Federal gives 20 to 25 July."],"geography":{"basis":"The protests and the facial-recognition deployment were at Jantar Mantar, New Delhi; the 25 were in Delhi's Tihar, Mandoli and Rohini prisons; the affidavit and order are in the Supreme Court of India (The Indian Express, IFF, The Federal).","court_countries":["IN"],"event_countries":["IN"],"affected_person_countries":["IN"]},"publication":{"basis":"Published under the 2026-09-15 charter as a contextual case in which an AI system made a consequential claim about identifiable people: a police facial-recognition system placed at least 25 people who were in prison at a protest, on a sworn list that the Supreme Court permitted police to act on. The finding rests on The Indian Express's records investigation (one chain, relayed by Outlook, IFF and The Federal) with the affidavit and order described as quoted; the 80 per cent threshold rests on IFF's 2022 RTI replies, which IFF says the Express also relies on. The listed people are not named.","reviewed_on":"2026-09-24"},"ai_involvement":{"basis":"Delhi Police's own affidavit, as quoted by The Indian Express, attributes the 2,873 identifications to its Facial Recognition System matching faces against the 'Crime Kundli' biometric database and criminal records; the Express's records check shows the system logged at least 25 people who were in custody. The affidavit was not inspected directly.","status":"reported"},"person_relations":["made_claim_about"]},"name":"Delhi: the police facial-recognition system logged at least 25 people who were in Tihar, Mandoli or Rohini jails as present at the Jantar Mantar student protests (20 to 26 July 2026), with timestamps on 24 to 26 July, on a sworn list of 2,873 persons with 'criminal antecedents' that the Supreme Court allowed police to register a fresh FIR against (Indian Express investigation, 4 September 2026)","summary":"In an affidavit of 17 August 2026 before the Supreme Court of India, Delhi Police said its Facial Recognition System (FRS) had identified 2,873 people with criminal antecedents at the Jantar Mantar protests of 20 to 26 July 2026 (2,402 through its 'Crime Kundli' biometric database and 471 through other criminal records). On 1 September the Court quashed the FIRs against the student protesters but let the police proceed against the 2,873; the police say any action will follow field verification. The Indian Express checked the 205 listed people facing murder, attempted-murder, rape or child-sexual-offence charges against police, prison and court records and found that at least 25 of them (17 accused of murder, four of rape (two or three under POCSO; the Express's narrative and its list differ), four of attempted murder) were lodged in Delhi's Tihar, Mandoli or Rohini prisons when the system logged them at the protest site with timestamps on 24, 25 and 26 July; some had been in custody for years. Delhi Police told the newspaper that further verification of the 2,873 was pending and, in its affidavit, that no action is taken solely on a facial-recognition result and that field verification follows each match. The police disclosed in 2022, in reply to a Right to Information request, that it treats a match with an 80 per cent similarity score as positive. Whether any of the 25 has since been named in the fresh FIR or visited for verification is not reported.","incidentDate":"2026-07-24","incidentEndDate":"2026-07-26","incidentKind":"bounded_series","incidentDatePrecision":"range","exposurePattern":"unknown","reportedDate":"2026-09-04","aiSystem":"Delhi Police Facial Recognition System (FRS): face-matching software run on protest footage against the police 'Crime Kundli' biometric database and other criminal records; the vendor and algorithm are not identified in the inspected reporting","aiProduct":"Unidentified facial recognition system","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["legal_harm"],"harmOutcomeSummary":"At least 25 people in judicial custody were recorded by the police facial-recognition system as present at a protest they could not have attended and were placed, in a sworn affidavit before the Supreme Court, on a list of 2,873 persons with criminal antecedents against whom the Court permitted a fresh FIR and police field verification; The Indian Express reports the misidentifications and the Internet Freedom Foundation describes the exposure to verification visits and prosecution as the harm. No arrest, FIR entry or verification visit of any of the 25 is reported in the inspected sources.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":0,"otherPeopleHarmedMin":25,"affectedCountStatus":"documented_minimum","affectedCountEvidence":"The Indian Express states 'at least 25 people on the list' were in jail when logged, and lists them (17 murder, four rape, four attempted murder). The other 2,668 entries were not checked, so the count is a documented minimum of 25.","victimAgeRange":"adult","jurisdiction":"IN","platformType":"other","outcomeType":"investigation_opened","outcomeStatus":"ongoing","primarySourceUrl":"https://indianexpress.com/article/express-exclusive/delhi-police-facial-recognition-jantar-mantar-protest-accused-in-jail-10862099/","primarySourceLabel":"The Indian Express, 4 September 2026: Police say cameras 'caught' murder, rape accused at CJP protest - some were actually in jail (Express Investigation)","firstPublishedAt":"2026-09-24T03:46:45.247544+00:00","updatedAt":"2026-09-30T01:17:35.08761+00:00","scopeVersion":"facts-v3","tags":["facial-recognition","india","delhi","protest-surveillance","supreme-court","police","false-match","made-claim-about","jantar-mantar"]},{"id":"2026-brazil-openai-report-user-detained-without-charge","caseFacts":{"claims":[{"id":"c1","status":"corroborated","evidence":[{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.","relation":"supports","source_id":"s1"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.","relation":"supports","source_id":"s3"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.","relation":"supports","source_id":"s4"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.","relation":"supports","source_id":"s2"}],"assertion":"Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.","causal_attribution":"Police account carried by two independent chains (BBC interviews; G1 with civil-police material)."},{"id":"c2","status":"reported","evidence":[{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): He was held in pre-trial detention for 54 days; on 13 August 2026 a court granted habeas corpus because, after more than 50 days, the police inquiry had not been concluded and no charge had been filed, and imposed an electronic ankle monitor and a ban on approaching or contacting his son and the child's mother; no charges had been brought and the investigation continues.","relation":"supports","source_id":"s1"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): He was held in pre-trial detention for 54 days; on 13 August 2026 a court granted habeas corpus because, after more than 50 days, the police inquiry had not been concluded and no charge had been filed, and imposed an electronic ankle monitor and a ban on approaching or contacting his son and the child's mother; no charges had been brought and the investigation continues.","relation":"supports","source_id":"s2"}],"assertion":"He was held in pre-trial detention for 54 days; on 13 August 2026 a court granted habeas corpus because, after more than 50 days, the police inquiry had not been concluded and no charge had been filed, and imposed an electronic ankle monitor and a ban on approaching or contacting his son and the child's mother; no charges had been brought and the investigation continues.","causal_attribution":"Court decision as reported by the BBC (single chain; the Spanish version is the same report)."},{"id":"c3","status":"disputed","evidence":[{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the reported messages were corroborated by material found during the arrest. The man denied intending harm and his lawyer disputed whether any concrete step had been taken.","relation":"supports","source_id":"s1"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the reported messages were corroborated by material found during the arrest. The man denied intending harm and his lawyer disputed whether any concrete step had been taken.","relation":"supports","source_id":"s2"}],"assertion":"Police say the reported messages were corroborated by material found during the arrest. The man denied intending harm and his lawyer disputed whether any concrete step had been taken.","causal_attribution":"Police and defence accounts as quoted by the BBC; the dispute concerns intent, not the detention."},{"id":"c4","status":"reported","evidence":[{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.","relation":"supports","source_id":"s1"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.","relation":"supports","source_id":"s2"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.","relation":"supports","source_id":"s3"}],"assertion":"OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.","causal_attribution":"OpenAI statements and investigators' accounts as reported; the absence of the replies is stated by investigators to the BBC."},{"id":"c5","status":"reported","evidence":[{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the conversations ran for about two months and an alert reached Brazilian authorities in June 2026, three days before the arrest.","relation":"supports","source_id":"s1"},{"locator":"Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the conversations ran for about two months and an alert reached Brazilian authorities in June 2026, three days before the arrest.","relation":"supports","source_id":"s2"}],"assertion":"Police say the conversations ran for about two months and an alert reached Brazilian authorities in June 2026, three days before the arrest.","causal_attribution":"BBC account (single chain)."}],"effects":[{"label":"held 54 days in pre-trial detention without charge after OpenAI reported his messages; released under electronic monitoring and a ban on contacting his son and the child's mother","claim_id":"c2","direction":"negative"},{"label":"police say the report let them arrest him a day before the date on which they believed the plan against the child would be carried out","claim_id":"c1","direction":"positive"}],"sources":[{"id":"s1","url":"https://www.bbc.com/portuguese/articles/c8dn1l7vr6ro","kind":"news_report","access":"read","language":"pt-br","translation_note":"Read live in Portuguese on 2026-09-23 (BBC News Brasil, Iara Diniz, 14 August 2026): messages seen by the BBC, police chief and defence-lawyer interviews, the habeas corpus release and precautionary measures. Translated by the reviewer.","independence_group":"bbc"},{"id":"s2","url":"https://www.bbc.com/mundo/articles/cxnvlg14e1p8o","kind":"news_report","access":"read","language":"es","translation_note":"Read live in Spanish on 2026-09-23 (BBC News Mundo, 16 September 2026): Spanish text produced by AI translation from an English-language BBC original and reviewed by a BBC journalist before publication (stated in the article); same reporting chain as s1, which is preferred where the two diverge. Translated by the reviewer.","independence_group":"bbc"},{"id":"s3","url":"https://g1.globo.com/tecnologia/noticia/2026/06/27/como-chatgpt-identificou-plano-de-pai-para-matar-o-filho-e-levou-alerta-para-autoridades.ghtml","kind":"news_report","access":"read","language":"pt-BR","translation_note":"Read live in Portuguese on 2026-09-23 (G1, 27 June 2026): arrest on 19 June, the OpenAI to FBI to Ministry of Justice to state police chain, message images from the civil police, OpenAI statement, police chief to TV Gazeta. Translated by the reviewer.","independence_group":"g1-globo"},{"id":"s4","url":"https://www.migalhas.com.br/quentes/459083/pai-e-preso-apos-planejar-morte-do-filho-no-chatgpt-veja-analise","kind":"news_report","access":"read","language":"pt","translation_note":"Read live in Portuguese on 2026-09-23 (Migalhas, 26 June 2026): summary attributed to an earlier G1/TV Gazeta report of 26 June, with lawyers' analysis; same Globo chain as s3. Translated by the reviewer.","independence_group":"g1-globo"}],"version":1,"ai_roles":["own_use","institutional_use"],"contexts":["justice","relationships","privacy"],"unknowns":["What ChatGPT replied during the two months of conversations (OpenAI did not provide the replies).","Whether charges will be brought after the forensic results and phone examination.","Whether he in fact contacted a hired killer, as he told ChatGPT.","The exact date OpenAI made its report to the FBI.","His identity (withheld by the BBC; not sought here)."],"geography":{"basis":"Arrest at his home in rural [identifying or operational detail withheld], [identifying or operational detail withheld] (BBC, G1, Migalhas); he lives there; habeas corpus granted by the Brazilian court handling the pre-trial detention (BBC: a Justiça). OpenAI's report went via the FBI in the United States, which is not an event location.","court_countries":["BR"],"event_countries":["BR"],"affected_person_countries":["BR"]},"publication":{"basis":"Published under the 2026-09-15 charter as a contextual case: an AI operator's report about a user, generated by ChatGPT safety systems, led to 54 days of pre-trial detention without charge, ended by habeas corpus. Two independent reporting chains (BBC; Globo) support the arrest and its cause; the release and its terms rest on the BBC. The police account that the report prevented a planned killing is preserved as context and intent is recorded as disputed. Nobody is named; the child is not identified. The charter states that this class needs no separate ruling.","reviewed_on":"2026-09-23"},"ai_involvement":{"basis":"He conversed with ChatGPT for about two months; OpenAI's safety systems flagged the conversations and the company reported them to the FBI, which passed them to Brazilian authorities (BBC, G1, with the police chief's account). OpenAI did not provide ChatGPT's replies, so what the chatbot said is unknown.","status":"supported"},"person_relations":["communicated_with","made_claim_about"]},"name":"Brazilian man detained without charge after OpenAI reported messages; he denies intending harm","summary":"BBC reporting describes a Brazilian man detained for 54 days after OpenAI reported messages to the FBI and Brazilian authorities. Police described a credible threat; he denied intending harm. A court released him without charges under protective conditions. The report did not include ChatGPT’s replies.","incidentDate":"2026-06-19","incidentEndDate":"2026-08-13","incidentKind":"bounded_series","incidentDatePrecision":"range","exposurePattern":"repeated_interactions","reportedDate":"2026-06-26","aiSystem":"ChatGPT (OpenAI); OpenAI's automated detection and human moderation systems, which generated the report","aiProduct":"ChatGPT and OpenAI moderation","aiCompany":"OpenAI","severity":"high","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["loss_of_liberty","legal_harm"],"harmOutcomeSummary":"The user was held in pre-trial detention for 54 days without charge on the strength of his ChatGPT messages as reported by OpenAI, then released under electronic monitoring and a no-contact order covering his son and the child's mother while the investigation continues. Police say the report was corroborated at the scene and prevented a planned killing; he and his lawyer dispute any intent. The consequence to him is undisputed; whether it was warranted is contested.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"One person, the ChatGPT user who was detained (BBC, G1). The child and the child's mother are not counted as harmed. Exact 1.","victimAgeRange":"adult","jurisdiction":"BR","platformType":"chatbot","outcomeType":"investigation_opened","outcomeStatus":"ongoing","primarySourceUrl":"https://www.bbc.com/portuguese/articles/c8dn1l7vr6ro","primarySourceLabel":"BBC News Brasil — report on the case","firstPublishedAt":"2026-09-23T03:45:57.312143+00:00","updatedAt":"2026-10-05T11:09:49.574842+00:00","scopeVersion":"facts-v3","tags":["chatgpt","openai","brazil","espirito-santo","law-enforcement-report","pre-trial-detention","fbi","institutional-response","communicated-with","made-claim-about"]}]}