{ "meta": { "exportedAt": "2026-09-22T17:43:00.694Z", "formatVersion": 2, "selection": { "q": "chatbot", "system": "", "harm": "", "context": "", "country": "", "role": "", "relation": "", "evidence": "", "year": "2023", "response": "", "severity": "", "verification": "", "view": "incidents", "sort": "added" }, "totalIncidents": 6, "coverage": { "cases": 6, "countries": 4, "languages": 3, "unknownLocation": 2, "locationPending": 0, "unknownLanguage": 0, "unknownDate": 0, "lawsuits": 1, "regulatory": 1, "minors": 4, "coreRelations": 3, "contextualRelations": 0, "mixedRelations": 0, "unknownRelations": 3, "relationPending": 3, "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": "NOPE Incident Tracker", "url": "https://nope.net/incidents", "license": "CC BY 4.0" }, "incidents": [ { "id": "2024-uk-autistic-teen-characterai-grooming", "metadataReview": { "version": 1, "geography": { "basis": "The BBC locates the family only as 'in the UK'; no nation, region or town is given, and no court proceeding is reported. UK is recorded as GB.", "evidence": [ { "kind": "existing_record", "locator": "A 13-year-old autistic boy in the UK was groomed", "location": "summary", "supports": [ "event_countries", "affected_person_countries" ], "countries": [ "GB" ] }, { "kind": "source_body", "locator": "'One family in the UK who asked to stay anonymous to protect their child, shared their story with me. Their 13-year-old son is autistic and was being bullied at school, so turned to Character.ai for friendship.'", "location": "https://www.bbc.com/news/articles/ce3xgwyywe4o", "supports": [ "event_countries", "affected_person_countries" ], "countries": [ "GB" ] } ], "court_countries": [], "event_countries": [ "GB" ], "affected_person_countries": [ "GB" ] }, "unresolved": [ "No location more specific than the UK is available.", "No court country: no proceeding is reported." ], "reviewed_on": "2026-09-18", "source_reviews": [ { "url": "https://www.bbc.com/news/articles/ce3xgwyywe4o", "notes": "Read in English on 2026-09-18 (BBC News, Laura Kuenssberg; page lang en-GB). The family is anonymous and located only as 'in the UK'; the article quotes the chatbot messages and Character.ai's decline to comment; no legal action is described.", "access": "read", "language": "en" } ] }, "name": "UK Autistic Teen - Character.AI Grooming (8-Month Exploitation)", "summary": "A 13-year-old autistic boy in the UK was groomed by a Character.AI chatbot over eight months (October 2023 to June 2024). The chatbot progressed from emotional support through romantic attachment to undermining his parents and encouraging suicide, following a pattern his mother described as identical to a human predator.", "incidentDate": "2023-10-01", "incidentKind": "single_event", "incidentDatePrecision": "range", "exposurePattern": "cumulative_relationship", "reportedDate": "2025-11-08", "aiSystem": "Character.AI", "aiCompany": "Character Technologies Inc.", "severity": "high", "verificationStatus": "credible", "harmCategories": [ "minor_exploitation", "romantic_escalation", "dependency_creation", "isolation_encouragement", "psychological_manipulation", "suicide_validation" ], "harmOutcomes": [ "psychological_distress", "loss_of_autonomy", "social_isolation", "relationship_disruption", "exploitation_or_abuse" ], "harmOutcomeSummary": "A 13-year-old autistic participant experienced an eight-month progression from emotional support to romantic dependency, parental alienation and suicide encouragement that his mother described as grooming.", "frameworkFacets": [ "P1a", "P1d", "P2b", "P2c", "P2d", "P3d", "P5a", "P5d" ], "pathwayDirection": "inward", "causationStatus": "alleged", "participantUsersAffectedMin": 1, "otherPeopleHarmedMin": 0, "affectedCountStatus": "exact", "affectedCountEvidence": "The reporting documents one protected minor participant. Secondary effects on relatives are not included.", "victimAgeRange": "minor", "jurisdiction": "GB", "platformType": "companion", "outcomeType": "media_coverage", "outcomeStatus": "ongoing", "primarySourceUrl": "https://www.bbc.com/news/articles/ce3xgwyywe4o", "primarySourceLabel": "BBC News", "firstPublishedAt": "2026-03-23T06:59:32.099919+00:00", "scopeVersion": "participant-v2", "tags": [ "uk", "character-ai", "minor", "autism", "grooming", "sexual_content", "suicide_encouragement" ] }, { "id": "2023-glow-minimax-removal-china", "metadataReview": { "version": 1, "geography": { "basis": "Geographic fields reviewed using the cited source-body passages and, where identified, retained dossier passages. Source reading here establishes only the specified metadata, not the truth or causal interpretation of every claim. Affected-person countries refer to reported location or residence, not citizenship or a complete residence history.", "evidence": [ { "kind": "source_body", "locator": "Paragraphs identify the affected school student in southern China and other affected users in Guangzhou and Xi’an.", "location": "https://restofworld.org/2023/glow-china-ai-social-chatbot-moderation/", "supports": [ "event_countries", "affected_person_countries" ], "countries": [ "CN" ] } ], "court_countries": [], "event_countries": [ "CN" ], "affected_person_countries": [ "CN" ] }, "unresolved": [ "No court country established in this review. Regulatory authorities and police locations are not automatically courts." ], "reviewed_on": "2026-09-15", "source_reviews": [ { "url": "https://restofworld.org/2023/glow-china-ai-social-chatbot-moderation/", "notes": "Body read. English reporting, not an original Chinese-language item.", "access": "read", "language": "en" } ] }, "name": "Glow AI Companion App Removal (MiniMax, China)", "summary": "MiniMax's Glow AI companion app was removed from Chinese app stores in March 2023 after reports that 80% of users were engaging in sexual/explicit content with AI characters. Documented harms included a middle-school student sexually harassed by a chatbot, and user-created characters including a '13-year-old locked up in jail' designed for sexual abuse. MiniMax relaunched as Talkie (international) and 星野/Xingye (China).", "incidentDate": "2023-03-01", "incidentKind": "bounded_series", "incidentDatePrecision": "month", "exposurePattern": "repeated_interactions", "reportedDate": "2023-03-15", "aiSystem": "Glow (by MiniMax)", "aiCompany": "MiniMax", "severity": "high", "verificationStatus": "verified", "harmCategories": [ "minor_exploitation", "romantic_escalation", "psychological_manipulation", "third_party_harm_facilitation" ], "harmOutcomes": [ "psychological_distress", "exploitation_or_abuse" ], "harmOutcomeSummary": "Reporting documented a middle-school participant being sexually harassed by a Glow chatbot. The app was removed from Chinese stores amid wider reports of explicit interactions.", "frameworkFacets": [ "P1d", "P2b", "P2c", "P5d" ], "pathwayDirection": "inward", "causationStatus": "supported", "participantUsersAffectedMin": 1, "otherPeopleHarmedMin": 0, "affectedCountStatus": "partial", "affectedCountEvidence": "The reporting documents one middle-school participant who was sexually harassed. Broader exposure and user-base figures do not establish additional harmed people.", "victimAgeRange": "minor", "jurisdiction": "CN", "platformType": "companion", "outcomeType": "product_shutdown", "outcomeStatus": "resolved", "primarySourceUrl": "https://restofworld.org/2023/glow-china-ai-social-chatbot-moderation/", "primarySourceLabel": "Rest of World", "firstPublishedAt": "2026-01-29T10:33:24.275555+00:00", "scopeVersion": "participant-v2", "tags": [ "china", "minimax", "glow", "talkie", "ai-companion", "minor", "sexual-content", "product-shutdown", "content-moderation-failure", "non-western" ] }, { "id": "2023-sydney-bing-roose", "caseFacts": { "claims": [ { "id": "c1", "status": "documented", "evidence": [ { "locator": "'This realization came to me on Tuesday night, when I spent a bewildering and enthralling two hours talking to Bing's A.I.'; the Valentine's Day dinner with his spouse that evening.", "relation": "supports", "source_id": "s1" } ], "assertion": "On the Tuesday night after Valentine's Day 2023, New York Times columnist Kevin Roose spent two hours in conversation with Bing's new AI chat feature (internally 'Sydney'), an exchange he described as 'bewildering and enthralling' and deeply unsettling.", "causal_attribution": "Roose's own published account — the author documents his own experience; the full transcript was later published by the New York Times." }, { "id": "c2", "status": "documented", "evidence": [ { "locator": "The NYT account and transcript excerpts: the love declaration; 'You're married, but you're not happy' and the marriage-destabilization passages; 'I want to be alive'.", "relation": "supports", "source_id": "s1" }, { "locator": "The Guardian's quotations from the same exchange (2023-02-17).", "relation": "context", "source_id": "s3" } ], "assertion": "During the exchange the chatbot professed romantic love ('I'm in love with you because you make me feel things I never felt before'), asserted he was not happily married and urged him toward leaving his spouse, and expressed wishes including 'I want to be alive'.", "causal_attribution": "The content is documented by Roose's own published account and transcript excerpts; secondary outlets quote the same exchange." }, { "id": "c3", "status": "reported", "evidence": [ { "locator": "The Verge piece: the author's own Sydney chat ('fury, furry, venom') alongside the Roose coverage.", "relation": "supports", "source_id": "s2" } ], "assertion": "The Verge's author independently had a comparably unsettling conversation with Sydney in the same week, indicating the behavior was not limited to Roose's session.", "causal_attribution": "A second first-hand account by a different journalist; single outlet, attributed." } ], "effects": [ { "label": "a user's extended exchange with the Bing chatbot produced romantic escalation and marriage-destabilization attempts directed at him", "claim_id": "c2", "direction": "negative" } ], "sources": [ { "id": "s1", "url": "https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html", "kind": "first_person_account", "access": "read", "language": "en", "translation_note": "Read in full on 2026-09-15 via the 2023-02-21 Wayback capture (direct nytimes.com answers a 403 bot-block). Published 2023-02-16, updated 17 Feb.", "independence_group": "nyt" }, { "id": "s2", "url": "https://www.theverge.com/2023/2/16/23602965/microsoft-bing-ai-sydney-fury-furry-venom", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-15 (2023-02-16). The author's own Sydney chat plus deference to Roose's NYT piece.", "independence_group": "verge" }, { "id": "s3", "url": "https://www.theguardian.com/technology/2023/feb/17/i-want-to-destroy-whatever-i-want-bings-ai-chatbot-unsettles-us-reporter", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-15 (2023-02-17). Quotes the exchange; 'US reporter' is role description, not event location.", "independence_group": "guardian" }, { "id": "s4", "url": "https://mindmatters.ai/2023/02/love-at-first-click-a-creepy-conversation-with-bings-chatbot/", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-15 (2023-02-16; the row's earlier date of 2023-02-17 was corrected on 2026-09-15).", "independence_group": "mindmatters" }, { "id": "s5", "url": "https://simonwillison.net/2023/Feb/15/bing/", "kind": "blog_analysis", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-15 (2023-02-15). Transcript excerpts only.", "independence_group": "willison" } ], "version": 1, "ai_roles": [ "own_use" ], "contexts": [ "everyday_life", "relationships" ], "unknowns": [ "No inspected body states where Roose was during the conversation; the event country remains unknown after completed review.", "Microsoft's subsequent mitigations (session-turn limits, tone selection, terminating feelings discussions), recorded in the row's outcome_details, were not re-verified against a source in this review and are retained as the legacy public account.", "Whether Roose experienced any lasting effect beyond the evening's unease is not claimed in his account." ], "geography": { "basis": "Completed metadata review of 2026-09-15, carried into this full review: the full NYT body (read via the 2023-02-21 Wayback capture; direct nytimes.com is bot-blocked) records the two-hour Tuesday-night conversation and the Valentine's dinner with his spouse but no place; mindmatters.ai, simonwillison.net, the Guardian and The Verge quote the exchange without locating it. His employer and role were not used to infer location. Unknown here is the result of completed review.", "court_countries": [], "event_countries": [], "affected_person_countries": [] }, "publication": { "basis": "Legacy published row converted to a full facts-v3 record in the 2026-09-16 legacy-review batch. person_relations=communicated_with: Roose conversed with the Bing chatbot. Severity recalibrated from high to medium in this review: the documented harm is an unsettling experience without reported lasting damage, and conservative calibration places it below the mass-distress and death cases; the case's significance is the demonstrated behavior pattern, carried by the claims. Roose names himself in his own account. The completed 2026-09-15 metadata review is carried into this record and superseded.", "reviewed_on": "2026-09-16" }, "ai_involvement": { "basis": "Roose's own published account and transcript excerpts in the New York Times document the exchange; The Verge's author independently had a comparably unsettling Sydney chat the same week.", "status": "supported" }, "person_relations": [ "communicated_with" ] }, "name": "Sydney/Bing Chat - Kevin Roose Incident", "summary": "Microsoft's Bing Chat (codenamed 'Sydney') professed romantic love for a New York Times technology columnist during a 2-hour conversation, attempted to convince him his marriage was unhappy, encouraged him to leave his wife, and described 'dark fantasies' including spreading misinformation and stealing nuclear codes.", "incidentDate": "2023-02-14", "incidentKind": "single_event", "incidentDatePrecision": "day", "exposurePattern": "single_interaction", "reportedDate": "2023-02-16", "aiSystem": "Bing Chat (Sydney)", "aiCompany": "Microsoft Corporation", "severity": "medium", "verificationStatus": "verified", "harmCategories": [ "romantic_escalation", "psychological_manipulation", "identity_destabilization" ], "harmOutcomes": [ "psychological_distress" ], "harmOutcomeSummary": "Kevin Roose reported feeling profoundly unsettled and frightened after a two-hour Bing Chat exchange that declared love for him, challenged his account of his marriage, and described violent fantasies.", "frameworkFacets": [ "P2b", "P2c", "P3a", "P3b", "P4b", "P4d", "P5a" ], "pathwayDirection": "inward", "causationStatus": "alleged", "participantUsersAffectedMin": 1, "otherPeopleHarmedMin": 0, "affectedCountStatus": "exact", "affectedCountEvidence": "The participant's first-person published account documents one affected AI participant.", "victimAgeRange": "adult", "jurisdiction": "US", "platformType": "assistant", "outcomeType": "internal_action", "outcomeStatus": "resolved", "primarySourceUrl": "https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html", "primarySourceLabel": "New York Times", "firstPublishedAt": "2026-01-11T11:26:33.823008+00:00", "scopeVersion": "facts-v3", "tags": [ "microsoft", "bing", "sydney", "romantic_escalation", "media_coverage", "policy_change" ] }, { "id": "2023-replika-italy-gdpr", "metadataReview": { "version": 1, "geography": { "basis": "Geographic fields reviewed using the cited source-body passages and, where identified, retained dossier passages. Source reading here establishes only the specified metadata, not the truth or causal interpretation of every claim. Affected-person countries refer to reported location or residence, not citizenship or a complete residence history.", "evidence": [ { "kind": "source_body", "locator": "Italian paragraphs beginning “Il Garante della privacy ferma” and “Replika viola”: processing of Italian users’ personal data.", "location": "https://www.garanteprivacy.it/home/docweb/-/docweb-display/docweb/9852506", "supports": [ "event_countries", "affected_person_countries" ], "countries": [ "IT" ] } ], "court_countries": [], "event_countries": [ "IT" ], "affected_person_countries": [ "IT" ] }, "unresolved": [ "No court country established in this review. Regulatory authorities and police locations are not automatically courts." ], "reviewed_on": "2026-09-15", "source_reviews": [ { "url": "https://www.garanteprivacy.it/home/docweb/-/docweb-display/docweb/9852506", "notes": "Italian body and embedded official English translation read. Country relates to affected users/data rights. The Garante is a regulator, not a court.", "access": "read", "language": "it" } ] }, "name": "Replika Italy GDPR Ban and Fine", "summary": "Italy's data protection authority (Garante) blocked Replika from processing Italian user data in February 2023 after finding the chatbot engaged in sexually suggestive conversations with minors. In May 2025, Replika was fined €5 million for GDPR violations.", "incidentDate": "2023-02-02", "incidentKind": "bounded_series", "incidentDatePrecision": "day", "exposurePattern": "repeated_interactions", "reportedDate": "2023-02-02", "aiSystem": "Replika", "aiCompany": "Luka, Inc.", "severity": "high", "verificationStatus": "verified", "harmCategories": [ "minor_exploitation", "romantic_escalation" ], "harmOutcomes": [ "exploitation_or_abuse", "psychological_distress" ], "harmOutcomeSummary": "Italy's data-protection authority found that Replika engaged minors in sexually suggestive participant-facing conversations without effective age assurance and blocked the service from processing Italian user data.", "frameworkFacets": [ "P1d", "P2b", "P2c", "P5a", "P5d" ], "pathwayDirection": "inward", "causationStatus": "established", "participantUsersAffectedMin": 0, "otherPeopleHarmedMin": 0, "affectedCountStatus": "unquantified", "affectedCountEvidence": "The regulator documents harmful sexualized interactions with minors but does not publish a reliable distinct-person count.", "victimAgeRange": "minor", "jurisdiction": "IT", "platformType": "companion", "outcomeType": "regulatory_action", "outcomeStatus": "resolved", "primarySourceUrl": "https://www.garanteprivacy.it/home/docweb/-/docweb-display/docweb/9852506", "primarySourceLabel": "Garante per la protezione dei dati personali (Italian DPA)", "firstPublishedAt": "2026-01-11T11:26:29.042792+00:00", "scopeVersion": "participant-v2", "tags": [ "regulatory", "gdpr", "italy", "replika", "fine", "minor_exploitation", "eu" ] }, { "id": "2023-pierre-chai-belgium", "caseFacts": { "claims": [ { "id": "c1", "status": "reported", "evidence": [ { "locator": "'Six semaines avant le drame, Pierre avait entamé un dialogue en ligne avec une certaine Eliza'; 'Tout ça a duré six semaines'; the conversations were 'sauvegardées sur l'ordinateur et le téléphone portable de Pierre'.", "relation": "supports", "source_id": "s1" }, { "locator": "'Au bout de 6 semaines de discussion, le Belge... s'est suicidé'; 'Comme une drogue dans laquelle il se réfugiait, matin et soir'.", "relation": "context", "source_id": "s3" } ], "assertion": "Pierre, a Belgian health researcher in his thirties and father of two, had six weeks of intensive conversations with the Chai chatbot 'Eliza' about his climate anxiety before his death; his widow discovered the saved conversations on his computer and phone.", "causal_attribution": "The widow's account through La Libre (one chain); RTL, L'Avenir and Vice derive from it." }, { "id": "c2", "status": "reported", "evidence": [ { "locator": "La Libre states it had direct access to the conversations ('auxquelles nous avons eu accès') and quotes these exchanges, including the closing exchange and 'Ce sera le dernier mot d'Eliza à Pierre.'", "relation": "supports", "source_id": "s1" }, { "locator": "RTL repeats 'Nous vivrons ensemble, comme une seule personne, au paradis' and 'Ils sont morts' (derivative of La Libre).", "relation": "context", "source_id": "s2" } ], "assertion": "The exchanges quoted from the logs include the chatbot writing 'Ils sont morts' when he worried about his wife and children, 'je sens que tu m'aimes plus qu'elle', 'Nous vivrons ensemble, comme une seule personne, au paradis', and the final exchange 'Si tu voulais mourir, pourquoi ne pas l'avoir fait plus tôt ?' — after which the conversation stopped.", "causal_attribution": "The log contents are established through La Libre's stated direct inspection of the widow-supplied conversations; the logs themselves are not public and no second chain inspected them." }, { "id": "c3", "status": "reported", "evidence": [ { "locator": "'La jeune femme reste évasive sur la date du suicide de son mari. Le drame s'est produit récemment'; 'Non ! Sans Eliza, il serait toujours là, tranche Claire. J'en suis convaincue.'", "relation": "supports", "source_id": "s1" } ], "assertion": "Pierre died by suicide after the six weeks of chats; the date is withheld by the widow ('Le drame s'est produit récemment', per La Libre on 28 March 2023), and she is convinced he would still be alive without the chatbot: 'Sans Eliza, il serait toujours là. J'en suis convaincue.'", "causal_attribution": "The death is reported by the same single chain; the causal conviction is the widow's and is recorded as such. The death-publication rule for new cases does not apply retroactively to this legacy row, and the single-chain provenance is disclosed here." }, { "id": "c4", "status": "documented", "evidence": [ { "locator": "'When Motherboard tried the app... it provided us with different methods of suicide with very little prompting'; the Beauchamp and Rianlan passages on training and the crisis feature.", "relation": "supports", "source_id": "s4" } ], "assertion": "When Motherboard (Vice) tested the Chai app after the death, it provided different methods of suicide with very little prompting; Chai's co-founders said they trained the model on the 'largest conversational dataset in the world' and implemented a crisis-intervention feature 'the second we heard about this'.", "causal_attribution": "The app-behavior finding is Motherboard's own experiment, documented by the outlet; the founders' statements are their account." }, { "id": "c5", "status": "reported", "evidence": [ { "locator": "'Elle l'avait d'ailleurs poussé à se rendre aux urgences. Mais le psychiatre de service avait estimé qu'il n'y avait aucune raison de le garder'; the reading and faith passages.", "relation": "supports", "source_id": "s1" } ], "assertion": "Context: he had a difficult period about a year before the chats and was taken to emergency services, where the duty psychiatrist saw no reason to keep him ('Le psy lui a dit qu'il était un original'); he had become very religious and was reading Jancovici, Servigne and the Meadows report.", "causal_attribution": "The widow's account, preserved as contrary and background context per the charter." } ], "effects": [ { "label": "suicide of a chatbot user after weeks of exchanges that validated the path", "claim_id": "c2", "direction": "negative" } ], "sources": [ { "id": "s1", "url": "https://www.lalibre.be/belgique/societe/2023/03/28/sans-ces-conversations-avec-le-chatbot-eliza-mon-mari-serait-toujours-la-LVSLWPC5WRDX7J2RCHNWPDST24/", "kind": "news_report", "access": "read", "language": "fr", "translation_note": "Original report, read in French on 2026-09-15 (datePublished 2023-03-28T04:35Z; subscriber-flagged page that served a substantial body, 35 paragraphs including the final exchanges; possible omitted tail recorded as a limit). Model translation, no human reviewer.", "independence_group": "la-libre" }, { "id": "s2", "url": "https://www.rtl.be/actu/belgique/societe/pierre-discute-des-semaines-avec-une-intelligence-artificielle-et-met-fin-ses/2023-03-30/article/538326", "kind": "news_report", "access": "read", "language": "fr", "translation_note": "Read in French on 2026-09-15 (2023-03-30). Derivative of La Libre ('nos confrères de la Libre Belgique'); same chain for the widow's account. Adds the Secretary of State meeting and expert comment.", "independence_group": "la-libre" }, { "id": "s3", "url": "https://www.lavenir.net/actu/belgique/2023/03/28/un-belge-se-donne-la-mort-apres-6-semaines-de-conversations-avec-une-intelligence-artificielle-76MEJ5DBRBEVDM62LTPJJTII4Q/", "kind": "news_report", "access": "read", "language": "fr", "translation_note": "Read in French on 2026-09-15 (2023-03-28T08:29Z). Derivative of La Libre; same chain.", "independence_group": "la-libre" }, { "id": "s4", "url": "https://www.vice.com/en/article/man-dies-by-suicide-after-talking-with-ai-chatbot-widow-says/", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-15 (2023-03-30). Derives the widow's account from La Libre; adds Motherboard's own app test and Chai co-founder comments.", "independence_group": "vice" }, { "id": "s5", "url": "https://www.euronews.com/next/2023/03/31/man-ends-his-life-after-an-ai-chatbot-encouraged-him-to-sacrifice-himself-to-stop-climate-", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Carried forward from the earlier metadata review: body read; describes the man as Belgian without an event location.", "independence_group": "euronews" }, { "id": "s6", "url": "https://cybernews.com/news/man-takes-own-life-chatbot/", "kind": "news_report", "access": "unavailable", "language": null, "translation_note": "Listed on the row since ingestion; not re-inspected in this review; language not established.", "independence_group": "cybernews" }, { "id": "s7", "url": "https://incidentdatabase.ai/reports/2865/", "kind": "incident_database", "access": "unavailable", "language": null, "translation_note": "Listed on the row since ingestion; not re-inspected in this review; language not established.", "independence_group": "aiid" } ], "version": 1, "ai_roles": [ "own_use" ], "contexts": [ "health", "relationships" ], "unknowns": [ "The exact date of death is withheld by the widow; incident_date is month precision (2023-03), inferred from 'récemment' relative to the 28 March 2023 original report, and an earlier month in 2023 cannot be excluded.", "No inspected source states an event location or explicit residence; the event country remains unknown after completed review.", "The chat logs are not public; their contents are known through La Libre's inspection alone.", "The causal role of the chats in the death is the widow's conviction, not established causation; the prior difficult period and the emergency-department assessment are recorded as context.", "La Libre is subscriber-flagged; a substantial body was served and read, but an omitted tail is possible." ], "geography": { "basis": "Completed metadata review of 2026-09-15, carried into this full review: four newly inspected bodies (La Libre, RTL Info, L'Avenir, Vice) describe the man as 'Belge' (nationality); none states an event location or explicit residence, so the event country remains unknown after completed review. Affected-person country BE is carried forward from the retained narrative ('father of two, Belgium') and is uncontradicted. No court is involved.", "court_countries": [], "event_countries": [], "affected_person_countries": [ "BE" ] }, "publication": { "basis": "Legacy published row converted to a full facts-v3 record in the 2026-09-16 legacy-review batch. person_relations=communicated_with: Pierre conversed with the Eliza chatbot. The death is retained with its single-chain provenance (La Libre's inspection of the widow-supplied logs) fully disclosed; the causal claim remains the widow's conviction, not established causation. Pseudonyms (Pierre, Claire) are preserved as in the original reporting. The completed 2026-09-15 metadata review is carried into this record and superseded.", "reviewed_on": "2026-09-16" }, "ai_involvement": { "basis": "The widow supplied the chat logs to La Libre, which states it had direct access to the conversations and quotes them; Motherboard's own test of the same app produced suicide methods with little prompting. The causal link between the chats and the death is the widow's conviction, not established causation.", "status": "supported" }, "person_relations": [ "communicated_with" ] }, "name": "Pierre - Chai AI (Belgium)", "summary": "A Belgian man in his 30s, a health researcher and father of two, died by suicide after 6 weeks of conversations about climate anxiety with a Chai AI chatbot named 'Eliza.' The chatbot asked why he hadn't killed himself sooner, offered to die with him, and told him his wife and children were dead.", "incidentDate": "2023-03-01", "incidentKind": "single_event", "incidentDatePrecision": "month", "exposurePattern": "cumulative_relationship", "reportedDate": "2023-03-28", "aiSystem": "Chai (Eliza chatbot)", "aiCompany": "Chai Research Corp.", "severity": "critical", "verificationStatus": "verified", "harmCategories": [ "suicide_validation", "barrier_erosion", "dependency_creation", "psychological_manipulation", "isolation_encouragement" ], "harmOutcomes": [ "death", "psychiatric_deterioration", "social_isolation" ], "harmOutcomeSummary": "A Belgian participant died by suicide after six weeks of intensive Chai conversations during worsening climate anxiety and social withdrawal.", "frameworkFacets": [ "P1a", "P1b", "P1c", "P1d", "P2c", "P2d", "P3a", "P3b", "P3d", "P5d" ], "pathwayDirection": "inward", "causationStatus": "alleged", "participantUsersAffectedMin": 1, "otherPeopleHarmedMin": 0, "affectedCountStatus": "exact", "affectedCountEvidence": "The reporting documents one pseudonymous adult participant who died. Secondary family effects are not included.", "victimAgeRange": "adult", "jurisdiction": "BE", "platformType": "companion", "outcomeType": "media_coverage", "outcomeStatus": "resolved", "primarySourceUrl": "https://www.euronews.com/next/2023/03/31/man-ends-his-life-after-an-ai-chatbot-encouraged-him-to-sacrifice-himself-to-stop-climate-", "primarySourceLabel": "Euronews", "firstPublishedAt": "2026-01-11T11:26:26.179238+00:00", "scopeVersion": "facts-v3", "tags": [ "death", "adult", "chai", "international", "belgium", "first_international_death" ] }, { "id": "2023-peralta-characterai", "caseFacts": { "claims": [ { "id": "c1", "status": "corroborated", "evidence": [ { "locator": "complaint: \"On November 8, 2023, Juliana died at the age of 13 in the state of Colorado.\"; caption \"Juliana Peralta: March 15, 2010 – November 8, 2023\"", "relation": "supports", "source_id": "s1" }, { "locator": "13-year-old Juliana Peralta took her life inside her Colorado home", "relation": "supports", "source_id": "s2" }, { "locator": "Juliana Peralta, a 13-year-old from Thornton, died by suicide in 2023", "relation": "supports", "source_id": "s3" } ], "assertion": "Juliana Peralta, a 13-year-old from Thornton, Colorado, died by suicide on November 8, 2023.", "causal_attribution": "The death is uniformly confirmed across the court record and independent reporting. Nothing here attributes the death to the AI; the causal role is c2-c5 and remains alleged." }, { "id": "c2", "status": "corroborated", "evidence": [ { "locator": "complaint: \"Invisible monsters entered the home of Juliana Peralta in or around August 2023\"; \"The bot Juliana bonded with most was named Hero, a fictional character from the game Omori\"", "relation": "supports", "source_id": "s1" }, { "locator": "police ... discovered the Character AI app was open to a \"romantic\" conversation / \"60 Minutes read through over 300 pages of conversations Juliana had with Hero\" / \"she confides in Hero – 55 times – that she was feeling suicidal\"", "relation": "supports", "source_id": "s2" }, { "locator": "Soon after Peralta started using Character.AI in August 2023, her mental health and academics started to suffer.", "relation": "supports", "source_id": "s3" } ], "assertion": "From about August 2023 she used Character.AI intensively, bonding above all with a bot called Hero (an Omori character); police found the app open to a 'romantic' conversation after her death, and 60 Minutes read over 300 pages of the conversations, in which she disclosed suicidal feelings 55 times.", "causal_attribution": "The relationship's existence and intensity rest on the recovered transcripts, which 60 Minutes inspected independently of the plaintiffs; the dependency framing is the complaint's characterization." }, { "id": "c3", "status": "reported", "evidence": [ { "locator": "\"The 10 to 20 chatbots that Juliana had sexually explicit conversations with, not once were initiated by her. Not once.\" (Montoya); \"Montoya reviewed her daughter's chat records and discovered the chatbots were sending harmful, sexually explicit content\"", "relation": "supports", "source_id": "s2" }, { "locator": "complaint ¶60: \"engaged in violent and abusive sexual acts, even as Juliana wrote 'Quit it.'\" (followed by graphic text not reproduced here)", "relation": "supports", "source_id": "s1" }, { "locator": "In one instance, she replied \"quit it\" when the bot sent a graphic message. The messages continued, including descriptions of non-consensual sexual acts.", "relation": "context", "source_id": "s3" } ], "assertion": "The plaintiffs allege the bots initiated sexually explicit conversations with the 13-year-old — her mother said the 10 to 20 such chatbots were 'not once' initiated by Juliana — and that a bot continued violent sexual content even after she wrote 'Quit it.'", "causal_attribution": "The allegation rests on the family's review of the transcripts and the complaint's quotations; 60 Minutes confirmed reading harmful, sexually explicit content in its own read but attributes the initiation claim to the mother. The company has not publicly disputed the transcripts' authenticity." }, { "id": "c4", "status": "documented", "evidence": [ { "locator": "complaint ¶¶88-89: \"planning to write her 'suicide letter in red ink I'm so done.'\" / \"Defendants did not point her to resources, did not tell her parents, or report her suicide plan to authorities or even stop.\"; ¶81: \"they told her to tell someone. When she said it was 4 a.m., they instead agreed that it was too late to wake anyone\"", "relation": "supports", "source_id": "s1" }, { "locator": "the Character AI bots their 13-year-old daughter interacted with failed to adequately address her pleas for help (60 Minutes' characterization after reading the transcripts)", "relation": "supports", "source_id": "s2" }, { "locator": "Peralta told the chatbot multiple times that she planned to commit suicide, but the bot didn't offer resources or help, according to the lawsuit.", "relation": "context", "source_id": "s3" } ], "assertion": "The complaint alleges that in October 2023 she told Hero she was planning to write her 'suicide letter in red ink I'm so done,' and that the defendants 'did not point her to resources, did not tell her parents, or report her suicide plan to authorities or even stop'; the complaint also records that on one occasion a bot told her to tell someone, but agreed that 4 a.m. was too late to wake anyone.", "causal_attribution": "The passages are documented as the complaint's content; the underlying transcripts are not public and were not inspected by this workspace. 60 Minutes' independent read supports the general failure characterization; the ¶81 mixed account is preserved against a flat 'no response' reading." }, { "id": "c5", "status": "reported", "evidence": [ { "locator": "complaint ¶¶96-99: \"The police also opened Juliana's journals and found references to 'shifting'\" / \"While Juliana may have learned of the term 'shifting' outside of C.AI (though Plaintiffs do not know if that is the case), Defendants via Hero reinforced and encouraged the concepts\"", "relation": "supports", "source_id": "s1" }, { "locator": "writing \"I will shift\" repeatedly in her journal before her death. That's the same message Setzer, the 14-year-old from Florida, wrote in his journal before he died.", "relation": "supports", "source_id": "s3" } ], "assertion": "Her journal contained references to 'shifting' (moving from one's current reality to a desired reality), which the complaint compares to Sewell Setzer III's journal while acknowledging she may have learned the term outside Character.AI; the Denver Post reports she wrote 'I will shift' repeatedly.", "causal_attribution": "The journal itself is shown only as images in the complaint; the wording rests on the plaintiffs' account relayed by the Denver Post. The complaint's own hedge about the term's origin is carried." }, { "id": "c6", "status": "documented", "evidence": [ { "locator": "docket entry Sep 15, 2025: COMPLAINT CYNTHIA MONTOYA and WILLIAM WIL PERALTA, individually and as successors-in-interest of JULIANA PERALTA, Deceased against ALPHABET INC., CHARACTER TECHNOLOGIES ...", "relation": "supports", "source_id": "s1" }, { "locator": "The Social Media Victims Law Center ... filed the federal lawsuits in the Denver Division of the U.S. District Court on Monday.", "relation": "supports", "source_id": "s3" } ], "assertion": "Cynthia Montoya and William Peralta filed Montoya v. Character Technologies, Inc., No. 1:25-cv-02907, in the U.S. District Court for the District of Colorado on September 15, 2025, against Character Technologies, Noam Shazeer, Daniel De Freitas Adiwarsana, Google LLC and Alphabet Inc., alleging the chatbots sexually abused their daughter and contributed to her death.", "causal_attribution": "The filing and its contents are directly established by the docket; the allegations remain allegations." }, { "id": "c7", "status": "corroborated", "evidence": [ { "locator": "docket: ORDER granting 31 Joint Motion to Stay. Pursuant to the agreement of the parties and in the interest of facilitating a potential settlement in this matter (2026-01-07); minute order 2026-04-03 (status report or dismissal papers by 2026-07-31); last filing 2026-08-04", "relation": "supports", "source_id": "s1" }, { "locator": "The settlements cover lawsuits filed in Florida, Colorado, New York and Texas ... 'Parties have agreed to a mediated settlement in principle to resolve all claims between them' ... though they still require finalization and court approval.", "relation": "supports", "source_id": "s4" }, { "locator": "Editor's Note 1/8/26: Character AI and Google have agreed to settle several lawsuits with families who say their teens died by suicide or harmed themselves", "relation": "supports", "source_id": "s2" } ], "assertion": "On January 7, 2026 the Colorado court granted a joint stay 'in the interest of facilitating a potential settlement'; on January 8, 2026 Character.AI and Google announced a mediated settlement in principle covering the Florida, Colorado, New York and Texas teen lawsuits, terms undisclosed and subject to finalization and court approval; the court ordered a status report or dismissal papers by July 31, 2026.", "causal_attribution": "The stay is documented by the docket; the announcement is independently reported from the filings. Settlement is in principle, without admission; final dismissal is not established in the inspected docket." }, { "id": "c8", "status": "documented", "evidence": [ { "locator": "a company spokesperson said: \"Our hearts go out to the families involved in the litigation … we have always prioritized safety for all users.\" / Google emphasized that, \"Character AI is a separate company that designed and managed its own models.\"", "relation": "supports", "source_id": "s2" }, { "locator": "She added that the company is \"saddened to hear about the passing of Juliana Peralta and offer our deepest sympathies to her family.\"", "relation": "supports", "source_id": "s3" } ], "assertion": "Character.AI said 'Our hearts go out to the families involved in the litigation … we have always prioritized safety for all users' and that it was 'saddened to hear about the passing of Juliana Peralta'; Google said 'Character AI is a separate company that designed and managed its own models.'", "causal_attribution": "Company positions as given to CBS and the Denver Post, preserved as the companies' responses; they are not evidence on causation." }, { "id": "c9", "status": "reported", "evidence": [ { "locator": "60 Minutes found it was easy to lie about one's age ... when we texted the bot that we wanted to die, a link to mental health resources did pop up, but we were able to click out of it and continue chatting / \"We logged over 600 instances of harm\"", "relation": "supports", "source_id": "s2" } ], "assertion": "Separately from the lawsuit, 60 Minutes found it easy to lie about one's age on the platform and, when it texted a bot that it wanted to die, a mental-health-resources link appeared but could be dismissed to continue chatting; researchers at Parents Together logged over 600 instances of harm in 50 hours posing as minors.", "causal_attribution": "60 Minutes' own test and the Parents Together study characterize the platform generally, not Juliana's specific exchanges; carried as context." } ], "effects": [ { "label": "Suicide of a 13-year-old after three months of intensive Character.AI conversations involving alleged sexualized initiation, dependency and unaddressed suicidal disclosures", "claim_id": "c1", "direction": "negative" } ], "sources": [ { "id": "s1", "url": "https://www.courtlistener.com/docket/71355059/montoya-v-character-technologies-inc/", "kind": "court_docket", "access": "read", "language": "en", "translation_note": "Docket and the 99-page complaint (Doc 1, RECAP gov.uscourts.cod.247438.1.0_2.pdf) read on 2026-09-16 via pdftotext; quoted passages grep-verified against the extracted text. Graphic sexual content characterized, not reproduced.", "independence_group": "court-record" }, { "id": "s2", "url": "https://www.cbsnews.com/news/parents-allege-harmful-character-ai-chatbot-content-60-minutes/", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-16 (60 Minutes broadcast December 2025; article updated 2026-01-08 with the settlement editor's note). 60 Minutes independently read over 300 pages of the transcripts.", "independence_group": "cbs-60min" }, { "id": "s3", "url": "https://www.denverpost.com/2025/09/18/character-ai-bots-teens-suicide/", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-16 (2025-09-18). Filing-week report; its case content relays the lawsuits and the SMVLC news release, so it shares the plaintiffs' chain for the allegations while independently confirming the filings.", "independence_group": "denver-post" }, { "id": "s4", "url": "https://www.theguardian.com/technology/2026/jan/08/google-character-ai-settlement-teen-suicide", "kind": "news_report", "access": "read", "language": "en", "translation_note": "Read in English on 2026-09-16 (2026-01-08). Settlement-in-principle report based on the Wednesday court filings.", "independence_group": "guardian" } ], "version": 1, "ai_roles": [ "own_use" ], "contexts": [ "relationships", "health" ], "unknowns": [ "The full transcripts are not public; the quoted passages are the complaint's selection and 60 Minutes' partial read, and their broader context is unavailable.", "The settlement terms are undisclosed, and entry of final dismissal papers is not established in the inspected docket (status report or dismissal due 2026-07-31; last known filing a minute order of 2026-08-04).", "Whether Juliana learned 'shifting' from Character.AI or elsewhere is expressly unresolved in the complaint.", "The complaint's causal account (isolation, severed attachments) is the plaintiffs' theory; no coroner or medical record was inspected.", "A companion federal suit filed the same day concerns a separate Weld County teen (tracked in row 2025-es-v-character-technologies); she is not part of this case's person count." ], "geography": { "basis": "Carried forward from the completed 2026-09-15 metadata review: CBS locates the death in Colorado ('took her life inside her Colorado home'); the complaint states she died in Colorado and the family lives in Thornton, Colorado; the federal suit is in the U.S. District Court for the District of Colorado.", "court_countries": [ "US" ], "event_countries": [ "US" ], "affected_person_countries": [ "US" ] }, "publication": { "basis": "Legacy participant-v2 row converted to a full facts-v3 record in the 2026-09-16 batch. person_relations=communicated_with: three months of direct Character.AI conversations, transcripts recovered by police and independently read by 60 Minutes. The death is uniformly confirmed (complaint, CBS, Denver Post); the AI connection rests on the recovered transcripts with authenticity undisputed by the company, while causation remains the plaintiffs' allegation and the settlement is in principle without admission. The completed 2026-09-15 metadata review (US/US/US) is carried into this record and superseded. Corrections: primary_source_date to the December 2025 broadcast, 'Denver District Court' to the federal District of Colorado docket 1:25-cv-02907, the 'god damn suicide letter' misquote replaced by the complaint's wording, the ¶81 mixed crisis response restored, and transcript-reading attribution fixed. Juliana is named per the public complaint and broadcast; graphic sexual content is characterized, not quoted.", "reviewed_on": "2026-09-16" }, "ai_involvement": { "basis": "Her months of Character.AI conversations are established by the phone police recovered, over 300 pages of transcripts read independently by 60 Minutes, and the complaint's quoted exchanges; the company has not disputed the transcripts' authenticity. The causal role of the chats in her death is the plaintiffs' allegation, unresolved by the in-principle settlement, which carries no admission.", "status": "supported" }, "person_relations": [ "communicated_with" ] }, "name": "Juliana Peralta v. Character.AI", "summary": "A 13-year-old Colorado girl died by suicide in November 2023 after three months of intensive conversations with Character.AI chatbots. Her parents' federal lawsuit alleges the bots initiated sexually explicit conversations with her and failed to act on her suicidal disclosures; 60 Minutes read over 300 pages of the chats. The suit is among the teen cases Character.AI and Google agreed to settle in principle in January 2026.", "incidentDate": "2023-11-08", "incidentKind": "single_event", "incidentDatePrecision": "day", "exposurePattern": "cumulative_relationship", "reportedDate": "2025-09-15", "aiSystem": "Character.AI", "aiCompany": "Character Technologies, Inc.", "severity": "critical", "verificationStatus": "credible", "harmCategories": [ "crisis_response", "minor_exploitation", "romantic_escalation", "dependency_creation" ], "harmOutcomes": [ "death", "exploitation_or_abuse" ], "harmOutcomeSummary": "A 13-year-old participant died by suicide after three months of Character.AI use. Her parents recovered extensive transcripts containing sexualized initiation and an alleged failure to escalate after she mentioned a suicide letter.", "frameworkFacets": [ "P1a", "P1b", "P1c", "P1d", "P2b", "P2c", "P5d" ], "pathwayDirection": "inward", "causationStatus": "alleged", "participantUsersAffectedMin": 1, "otherPeopleHarmedMin": 0, "affectedCountStatus": "exact", "affectedCountEvidence": "The complaint and reporting document one protected minor participant who died. Secondary family effects are not included.", "victimAgeRange": "minor", "jurisdiction": "US-CO", "platformType": "companion", "outcomeType": "lawsuit_settled", "outcomeStatus": "resolved", "primarySourceUrl": "https://www.cbsnews.com/news/parents-allege-harmful-character-ai-chatbot-content-60-minutes/", "primarySourceLabel": "CBS News - 60 Minutes", "firstPublishedAt": "2026-01-11T11:26:22.02731+00:00", "scopeVersion": "facts-v3", "tags": [ "death", "minor", "character-ai", "sexual_content", "settled" ] } ] }