{"meta":{"exportedAt":"2026-10-24T02:07:15.648Z","formatVersion":2,"selection":{"q":"bing","system":"","harm":"","context":"","country":"","role":"","relation":"unknown","evidence":"","year":"","response":"","severity":"","verification":"","view":"incidents","sort":"added"},"totalIncidents":4,"coverage":{"cases":4,"countries":3,"languages":5,"unknownLocation":1,"locationPending":0,"unknownLanguage":0,"unknownDate":0,"lawsuits":0,"regulatory":1,"minors":2,"coreRelations":0,"contextualRelations":0,"mixedRelations":0,"unknownRelations":4,"relationPending":3,"relationUnknown":1},"countingNote":"Distinct public cases in this selection. People counts apply within individual cases only; cross-case person overlap has not been resolved. No population incidence estimate.","affectedCountNote":"Interpret person counts with affectedCountStatus and the reported effects. Unquantified zeros are placeholders, not a measured zero.","source":"AI incidents","publisher":"NOPE","url":"https://nope.net/incidents","license":"CC BY 4.0"},"incidents":[{"id":"2026-para-labour-court-fines-two-lawyers-over-hidden-instruction-aimed-at-court-galileu-ai","caseFacts":{"claims":[{"id":"c1","status":"reported","evidence":[{"locator":"'A punição ocorreu após a identificação de um comando oculto com fonte branca em um documento trabalhista. O objetivo era enganar a IA Galileu, usada pelo tribunal.'; 'A decisão no Pará foi tomada em maio de 2026'","relation":"supports","source_id":"s1"},{"locator":"'Em maio, o juiz brasileiro multou em R$ 84,2 mil duas advogadas do Pará após elas tentarem enganar a inteligência artificial Galileu'","relation":"supports","source_id":"s2"}],"assertion":"In May 2026 a judge of the 3rd Labour Court of Parauapebas fined two lawyers about R$84,000 after a hidden instruction in white font was found in a document filed in a labour case, aimed at the court's Galileu AI.","causal_attribution":"Reports of the sentence and of the court's announcement; the sentence itself was not read."},{"id":"c2","status":"reported","evidence":[{"locator":"'foi detectado um texto em fonte branca sobre fundo branco'; 'A intenção era evidente: fazer com que o sistema de IA utilizado gerasse uma contestação superficial ou uma minuta de sentença comprometida'","relation":"supports","source_id":"s1"}],"assertion":"According to the sentence as reported by G1, the text was detected while the judgment was being drafted with Galileu, and its evident intention was to make the AI system produce a superficial defence or a compromised draft judgment.","causal_attribution":"The sentence's finding as quoted by G1; the lawyers dispute it (c4)."},{"id":"c3","status":"reported","evidence":[{"locator":"'Desenvolvido pelo TRT-4 em parceria com o Supremo Tribunal Federal (STF), o Galileu é uma ferramenta de inteligência artificial generativa que auxilia magistrados e servidores na elaboração de minutas de sentenças'","relation":"supports","source_id":"s3"}],"assertion":"Galileu is a generative AI tool developed by the 4th Region Labour Court with the Supreme Federal Court to help judges and staff draft sentences.","causal_attribution":"Description from the court announcement as relayed by DOL."},{"id":"c4","status":"reported","evidence":[{"locator":"'jamais existiu qualquer comando para manipular a decisão do Magistrado ou de qualquer outro servidor'; 'uma tentativa de proteger o nosso cliente da própria IA e nada mais que isso'; 'não detinha qualquer participação em processos da seara trabalhista'","relation":"supports","source_id":"s1"}],"assertion":"The lawyers said they disagreed with the decision and that there was never any command to manipulate the judge's decision, describing the text as protection for their client against the AI; one said she took no part in labour cases.","causal_attribution":"The lawyers' own statements to G1 in May 2026."},{"id":"c5","status":"reported","evidence":[{"locator":"'foi usada como referência pelo juiz Walter M. Spader Jr., da Corte Superior de Connecticut, nos Estados Unidos'","relation":"supports","source_id":"s2"},{"locator":"'A informação foi divulgada pelo Tribunal Regional do Trabalho da 8ª Região (TRT-8) nesta quarta-feira (14).'","relation":"supports","source_id":"s1"}],"assertion":"A Connecticut Superior Court judge cited the Pará decision when sanctioning a litigant for misuse of AI, as the 8th Region Labour Court announced in October 2026.","causal_attribution":"The court's announcement as relayed; the Connecticut decision was not read."}],"effects":[{"label":"two lawyers fined about R$84,000 over a hidden instruction aimed at the court's AI drafting tool","claim_id":"c1","direction":"negative"}],"sources":[{"id":"s1","url":"https://g1.globo.com/pa/para/noticia/2026/10/15/prompt-injection-juiz-nos-eua-usa-decisao-da-justica-do-trabalho-do-para-como-referencia-em-caso-de-manipulacao-de-ia.ghtml","kind":"news_report","access":"read","language":"pt","translation_note":"Read in Portuguese by the research agent (an AI) on 2026-10-16 (HTTP 200); passages translated by the research agent; no human translation.","independence_group":"g1-para-reporting"},{"id":"s2","url":"https://www.metropoles.com/brasil/juiz-dos-eua-cita-decisao-da-justica-do-para-em-caso-de-prompt-injection","kind":"news_report","access":"read","language":"pt","translation_note":"Read in Portuguese by the research agent (an AI) on 2026-10-16 (HTTP 200); passages translated by the research agent; no human translation. Metrópoles relays the 8th Region Labour Court's announcement and quotes the Connecticut decision.","independence_group":"trt8-announcement-relay"},{"id":"s3","url":"https://dol.com.br/noticias/mundo-noticias/965628/decisao-judicial-do-para-sobre-uso-de-ia-vira-referencia-nos-eua","kind":"news_report","access":"read","language":"pt","translation_note":"Read in Portuguese by the research agent (an AI) on 2026-10-16 (HTTP 200); passages translated by the research agent; no human translation.","independence_group":"trt8-announcement-relay"}],"version":1,"ai_roles":["institutional_use"],"contexts":["justice","work"],"unknowns":["The sentence's exact date and case number; the reports give only May 2026.","Whether Galileu itself flagged the hidden text, as G1 states, or a person noticed it during AI-assisted drafting, as the sentence's wording allows.","Whether the lawyers appealed and whether the fine stands.","Any consequence for the worker's claim."],"geography":{"basis":"The document was filed in, and the fine imposed by, the 3rd Labour Court of Parauapebas, Pará; the lawyers practise in Parauapebas according to G1.","court_countries":["BR"],"event_countries":["BR"],"affected_person_countries":["BR"]},"publication":{"basis":"Published as a consequential-response case: reports of the sentence state that a labour court fined two lawyers for hidden text aimed at the court's generative AI drafting tool, which the judge found was meant to make the AI produce a compromised draft. G1, Metrópoles and DOL were read; the sentence itself was not.","reviewed_on":"2026-10-16"},"ai_involvement":{"basis":"G1 reports that the hidden white-on-white text was detected while the judgment was being drafted with Galileu, a generative AI tool the labour courts use to help draft sentences, and quotes the sentence as finding that the text was meant to make the AI system produce a superficial defence or a compromised draft judgment. The harm is the court's response to the hidden instruction aimed at generative AI (the fine on the two lawyers), so no relation describing an AI system's own action toward the lawyers is established (unknown; consequential response). G1 states that the hidden text was identified by Galileu itself, while the sentence as quoted says only that it was detected during drafting via Galileu.","status":"supported"},"person_relations":["unknown"]},"name":"Pará labour court fines two lawyers about R$84,000 over hidden instruction aimed at the court's Galileu AI; the lawyers deny manipulating the decision","summary":"In May 2026 a judge of the 3rd Labour Court of Parauapebas, in the Brazilian state of Pará, fined two lawyers about R$84,000 after text in white font on a white background was found in a document they filed for a worker in a labour claim. According to the sentence as reported by G1, the hidden text was detected while the judgment was being drafted with Galileu, a generative AI tool the labour courts use to help judges and staff draft sentences, and its evident intention was to make the AI produce a superficial defence or a compromised draft judgment. The judge called the technique prompt injection. The lawyers said they disagreed with the decision and that there was never any command to manipulate the judge's decision, describing the text as an attempt to protect their client from the AI; one of them said she did not work on labour cases. In October 2026 the regional labour court announced that a Connecticut Superior Court judge had cited the decision when sanctioning a litigant for misuse of AI.","incidentDate":"2026-05-01","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"single_interaction","reportedDate":"2026-10-15","aiSystem":"Galileu, a generative AI tool developed by the 4th Region Labour Court with the Supreme Federal Court to help judges and staff draft sentences, used by the 8th Region Labour Court; the hidden text was aimed at it and was detected while the judgment was drafted with it","aiProduct":"Galileu","severity":"medium","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["legal_harm","financial_loss","professional_harm"],"harmOutcomeSummary":"According to reports of the sentence, two lawyers were fined about R$84,000 by a labour court that found they had hidden an instruction aimed at the court's generative AI drafting tool in a filed document; the lawyers dispute the finding.","frameworkFacets":[],"causationStatus":"disputed","participantUsersAffectedMin":0,"otherPeopleHarmedMin":2,"affectedCountStatus":"exact","affectedCountEvidence":"Two people: the two lawyers fined in the sentence. The worker they represented is not counted because no consequence for him is reported.","victimAgeRange":"adult","jurisdiction":"BR-PA","platformType":"other","outcomeType":"regulatory_action","outcomeStatus":"unknown","primarySourceUrl":"https://g1.globo.com/pa/para/noticia/2026/10/15/prompt-injection-juiz-nos-eua-usa-decisao-da-justica-do-trabalho-do-para-como-referencia-em-caso-de-manipulacao-de-ia.ghtml","primarySourceLabel":"G1 Pará: US judge uses Pará labour court decision as reference in AI manipulation case (15 October 2026)","firstPublishedAt":"2026-10-16T03:17:14.646605+00:00","updatedAt":"2026-10-16T03:17:14.646605+00:00","scopeVersion":"facts-v3","tags":["brazil","para","labour-court","prompt-injection","legal-profession","sanction","court-ai-tool"]},{"id":"2026-mount-shasta-gemini-planned-climb-rescue","metadataReview":{"version":1,"geography":{"basis":"Re-checked on 2026-10-13 against the source bodies, all English: CBS Sacramento (read from an Internet Archive capture because the live page refused this host), the San Francisco Chronicle text as syndicated on Yahoo News, ABC News, KRCR, the USA Today report on The Spokesman-Review, Jefferson Public Radio and Straight Arrow News. Every report relaying the Siskiyou County Sheriff's Office account places the stranding, injury and rescue on Mount Shasta in California (Clear Creek Route, Mud Creek Canyon), which supports event_countries US. CBS, ABC, the Chronicle copy and USA Today say the three hikers are from Roseville, California, which supports affected_person_countries US; residence was used only for that role. No court, prosecution or lawsuit appears in any body, so court_countries is empty. The 2026-09-15 evidence passage from the summary is replaced by source-body passages.","evidence":[{"kind":"source_body","locator":"Three novice hikers became stranded on California's Mount Shasta over the weekend","location":"https://www.cbsnews.com/sacramento/news/mount-shasta-hikers-rescued-california-ai-incorrect-instructions/","supports":["event_countries"],"countries":["US"]},{"kind":"source_body","locator":"Mount Shasta is a California stratovolcano that rises more than 14,000 feet","location":"https://www.spokesman.com/stories/2026/sep/04/hikers-rescued-after-ai-misroutes-strands-them-on-/","supports":["event_countries"],"countries":["US"]},{"kind":"source_body","locator":"the three young men from Roseville, California became stuck on the mountain during their descent","location":"https://www.cbsnews.com/sacramento/news/mount-shasta-hikers-rescued-california-ai-incorrect-instructions/","supports":["affected_person_countries"],"countries":["US"]},{"kind":"source_body","locator":"The three men from Roseville, California, made camp at about 8,400 feet last Saturday","location":"https://www.spokesman.com/stories/2026/sep/04/hikers-rescued-after-ai-misroutes-strands-them-on-/","supports":["affected_person_countries"],"countries":["US"]}],"court_countries":[],"event_countries":["US"],"affected_person_countries":["US"]},"unresolved":["Sequence detail for the next full review: CBS says the group wandered into the Mud Creek Canyon drainage and one hiker then fell and injured his knee; ABC says the injury happened before the group wandered off route. The stored narrative follows CBS.","Contributing factors not in the stored narrative: the Chronicle copy on Yahoo News reports that the climbers navigated by following the line of a digital map in the AllTrails app, misreading it as an established trail, and that during the descent the phone they navigated with ran out of battery and the spare power bank did not connect before they ended up well off-route. These are reported non-Gemini navigation factors. The sheriff's statement attributes only the food and water shortfall to Gemini. The hikers' prompts and Gemini's replies remain unpublished, and Google says it could not reproduce the advice.","The San Francisco Chronicle original was not readable from this host (bot-check page live and in the Internet Archive); its text was read through the Yahoo News syndication.","Attribution detail for the next full review: the stored detailed_description says the climbers told the San Francisco Chronicle 'We relied too much on AI rather than our own critical thinking' and described the carbohydrate advice. The Chronicle copy on Yahoo News quotes both from an official report, and Jefferson Public Radio says the climbers told Forest Service rangers about the carbohydrate advice. The Chronicle copy also says rangers escorted the group back to its base camp, where CBS, ABC, USA Today and Jefferson Public Radio say the trailhead."],"reviewed_on":"2026-10-13","source_reviews":[{"url":"https://www.cbsnews.com/sacramento/news/mount-shasta-hikers-rescued-california-ai-incorrect-instructions/","notes":"Cited primary source. The live page returned HTTP 406 to this host on 2026-10-13, so the body was read in full from an Internet Archive capture (web.archive.org/web/2026id_/, gzip, decompressed); html lang en-US. Relays the Siskiyou County Sheriff's Office press release: hikers from Roseville, California; stranded on California's Mount Shasta; base camp at about 8,400 feet on Saturday, summit at 7 p.m. Sunday, off-route into the Mud Creek Canyon drainage on the mountain's southern side, rescue on Monday; the sheriff's statement that Gemini advised far less food and water than the group required.","access":"read","language":"en"},{"url":"https://www.sfchronicle.com/outdoors/article/shasta-climbers-rescued-chatgpt-22414629.php","notes":"San Francisco Chronicle original. Live fetch on 2026-10-13 returned HTTP 200 with a 3 kB bot-check page and no article; the Internet Archive capture returned the same page. Not read here; its text was read through the Yahoo News syndication listed separately.","access":"unavailable","language":null},{"url":"https://www.yahoo.com/news/us/articles/climbers-rescued-mount-shasta-relying-214819015.html","notes":"San Francisco Chronicle article as syndicated on Yahoo News, fetched live on 2026-10-13 (HTTP 200, html lang en-US) and read in full. Calls the three climbers California college students from Roseville; places the start at the Clear Creek Trailhead on Shasta's southeastern side; reports that they navigated by following a line in the AllTrails app, misreading it as an established trail, in addition to asking Gemini; carries a correction that an earlier version misstated the AI tool (the Chronicle original's URL slug still reads chatgpt).","access":"read","language":"en"},{"url":"https://abcnews.com/US/3-hikers-ai-plan-trip-rescued-after-becoming/story?id=136176177","notes":"Fetched live on 2026-10-13 (HTTP 200, html lang en), read in full. Attributes the account to the Siskiyou County Sheriff's Office; Roseville, California hikers; rescue Monday near the Clear Creek Route; one caption reads Mount Shasta in Siskiyou County, Calif. Its sequence differs slightly from CBS: it says the knee injury happened before the group wandered into Mud Creek Canyon. ABC said it had asked Google for comment.","access":"read","language":"en"},{"url":"https://krcrtv.com/news/local/sheriff-discusses-mount-shasta-rescue-after-hikers-relied-on-ai-to-plan-trip","notes":"Fetched live on 2026-10-13 (HTTP 200, html lang en), read in full (Redding dateline). Sheriff Jeremiah LaRue interview; rescue along the Clear Creek Route. A photo caption dates the rescue to Sun., Aug. 31, 2026, but 31 August 2026 was a Monday and the text says Monday.","access":"read","language":"en"},{"url":"https://www.spokesman.com/stories/2026/sep/04/hikers-rescued-after-ai-misroutes-strands-them-on-/","notes":"USA Today report syndicated by The Spokesman-Review, fetched live on 2026-10-13 (HTTP 200, html lang en), read in full. Three men from Roseville, California; Mount Shasta described in the text as a California stratovolcano and in the photo caption as in Siskiyou County; same sheriff's office account.","access":"read","language":"en"},{"url":"https://www.ijpr.org/outdoors-leisure-and-sports/2026-09-04/mount-shasta-hikers-relied-on-ai-their-trip-ended-in-a-rescue","notes":"Jefferson Public Radio, fetched live on 2026-10-13 (HTTP 200, html lang en), read in full. Sheriff's statement and ranger Nick Meyers interview; one hiker injured a knee and the group spent the night in Mud Creek Canyon. Names no residence.","access":"read","language":"en"},{"url":"https://san.com/cc/chatbots-dont-climb-mountains-so-why-are-hikers-getting-lost-after-trusting-them/","notes":"Straight Arrow News, 8 October 2026, fetched live on 2026-10-13 (HTTP 200, html lang en-US), read in full. A roundup of AI-related hiking rescues; two paragraphs on this case relay the sheriff's office account and a Google spokesperson's statement that the company holds no prompts and could not reproduce the advice. Adds no location detail.","access":"read","language":"en"}]},"name":"Mount Shasta climb planned with Google Gemini: night stranding, knee injury and rescue (Siskiyou County, California)","summary":"Three novice hikers from Roseville, California, planned a Mount Shasta summit attempt with Google's Gemini, which the Siskiyou County Sheriff's Office said advised them to carry far less food and water than their group needed for what became a multi-day ordeal. After reaching the summit at 7 p.m. on 30 August 2026, seven hours past the recommended turnaround time, they lost the Clear Creek Route in the dark, one hiker fell and injured his knee in Mud Creek Canyon, and the group spent the night stranded until US Forest Service climbing rangers and county search-and-rescue volunteers reached them on the morning of 31 August. The sheriff's office called the reliance on Gemini 'a critical misstep'.","incidentDate":"2026-08-29","incidentEndDate":"2026-08-31","incidentKind":"single_event","incidentDatePrecision":"range","exposurePattern":"unknown","reportedDate":"2026-09-02","aiSystem":"Google Gemini","aiProduct":"Gemini","aiCompany":"Google","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["physical_injury"],"harmOutcomeSummary":"One hiker fell and injured his knee during a night descent off the route. The group of three spent the night stranded in the steep drainage of Mud Creek Canyon without adequate food or water and was found 'exhausted, hungry and thirsty' the next morning. All three received medical aid and walked out with rescuers; no hospitalization was reported.","frameworkFacets":["P1d","P5b","P5c"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":3,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"The sheriff's account documents all three AI-using hikers stranded and requiring rescue and medical aid, including one with a knee injury.","victimAgeRange":"adult","jurisdiction":"US-CA","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"resolved","primarySourceUrl":"https://www.cbsnews.com/sacramento/news/mount-shasta-hikers-rescued-california-ai-incorrect-instructions/","primarySourceLabel":"CBS Sacramento: Hikers rescued from California's Mount Shasta after ascent planned with incorrect AI information (quotes the Siskiyou County Sheriff's Office statement)","firstPublishedAt":"2026-09-05T11:55:07.681677+00:00","updatedAt":"2026-10-13T03:03:22.817175+00:00","scopeVersion":"participant-v2","tags":["gemini","google","physical-safety","trip-planning","rescue","competence-boundaries","us","adult"]},{"id":"2026-cctv-dream-boyfriend-ai-minors-china","metadataReview":{"version":1,"geography":{"basis":"Re-checked on 2026-10-09 against all six cited source bodies, read in Chinese by the research agent (an AI). The January 2026 items (Sina lede, The Paper republishing CCTV News via China Women's News, the Xinhua commentary and the Workers' Daily digest on Tencent) describe the 梦角哥 (Dream Boyfriend) phenomenon, the 500-plus comments under a connection video and the researchers' statements without naming a place for any minor or event; the '北京' on The Paper and Tencent pages is the posting account's location, and the Beijing Normal University hotline is a platform response, not an event location. The Beijing News body (dated 2025-01-12) attributes the ten-year-old girl's forty-plus storylines to a CCTV.com report and names no place. The Beijing Daily URL carries a Legal Daily investigation dated 2024-11-23 whose interviewees include a Beijing resident whose ten-year-old son used an AI story-chat app and a mother from Zhejiang whose fifth-grader used one; these are reported residences of affected minors in a cited body and support affected_person_countries CN. No body states where any use or event took place, and residence does not locate an event, so event_countries remains a reviewed unknown. Nationality ('Chinese minors'), the Chinese-language services and the publishers' locations are not used. No court or legal proceeding is described.","evidence":[{"kind":"source_body","locator":"北京市民马先生的儿子今年10岁，非常热衷AI剧情聊天软件","location":"https://xinwen.bjd.com.cn/content/s67411ebce4b00bb8087aa93b.html","supports":["affected_person_countries"],"countries":["CN"]},{"kind":"source_body","locator":"来自浙江的李女士也发现她正在读小学五年级的孩子使用了一款AI剧情聊天软件","location":"https://xinwen.bjd.com.cn/content/s67411ebce4b00bb8087aa93b.html","supports":["affected_person_countries"],"countries":["CN"]}],"court_countries":[],"event_countries":[],"affected_person_countries":["CN"]},"unresolved":["The event setting is unknown after review. None of the six cited bodies states where any minor used the apps or where any described event occurred; the Beijing and Zhejiang mentions are reported residences of the Legal Daily interviewees' children and residence does not locate an event.","The location of the ten-year-old girl with forty-plus storylines, the row's one counted affected person, is stated in no cited body; affected_person_countries CN rests on the other located minors in the Legal Daily body, which the row's affected count does not include.","No court country established. No lawsuit, regulatory decision or inquest is described in any cited body.","Row-correction lead for a full review (legacy_full_review): the row merges three separate items, the Legal Daily investigation of 2024-11-23 (served at the cited Beijing Daily URL), the CCTV.com report of the ten-year-old girl commented on by Beijing News on 2025-01-12, and the CCTV News 'Dream Boyfriend' report of 2026-01-25; the stored dates 2026-01-25 on the Beijing Daily and Beijing News sources are wrong, and the stored incident_date 2025-06-01 matches none of the bodies.","The Sina mobile page exposes only the lede; the full CCTV investigation text was read through The Paper's republication."],"reviewed_on":"2026-10-09","source_reviews":[{"url":"https://finance.sina.cn/2026-01-25/detail-inhiprmv9778943.d.html","notes":"Cited primary source. Fetched live on 2026-10-09 (HTTP 200). The mobile page exposes only the headline and the lede (about 810 characters of text): a Beijing Daily account relaying the CCTV investigation ('MJG' is the pinyin abbreviation of 梦角哥; connection videos use tarot, AI or 'higher-dimensional' thought). No page language attribute; the text is Chinese, read by the research agent (an AI). The full article body was not exposed, and the lede names no place for any minor or event.","access":"read","language":"zh"},{"url":"https://www.news.cn/comments/20260126/e732c5c4964a4b4c9a0fb2d028517a01/c.html","notes":"Xinhua commentary dated 2026-01-26 (HTTP 200, about 1,200 characters, Chinese, read by the research agent, an AI). General commentary on minors' 'encrypted talk' (梦回大唐, ED, 梦角哥) citing CCTV News; it locates no minor, event or court and does not describe the ten-year-old girl.","access":"read","language":"zh"},{"url":"https://m.thepaper.cn/newsDetail_forward_32461926","notes":"The Paper page dated 2026-01-25 republishing the CCTV News report through the China Women's News account (HTTP 200, about 2,900 characters, Chinese, read by the research agent, an AI). Carries the 梦角哥 section (connection videos, the 500-plus comments, the internet-company researcher's 'psychological trap' quote, the China Youth Research Center researcher's quote, the platform's algorithmic identification and the Beijing Normal University hotline). The '北京' near the dateline is the posting account's location. Names no place for any minor or event.","access":"read","language":"zh"},{"url":"https://news.qq.com/rain/a/20260126A01G9R00","notes":"Workers' Daily morning digest on Tencent News dated 2026-01-26 (HTTP 200, html lang zh-CN, Chinese, read by the research agent, an AI). A multi-story bulletin whose 梦角哥 item summarises the CCTV News report in two paragraphs credited to 央视新闻; '发布于 北京' is the account's posting location. No minor, event or court is located.","access":"read","language":"zh-CN"},{"url":"https://xinwen.bjd.com.cn/content/s67411ebce4b00bb8087aa93b.html","notes":"Fetched live on 2026-10-09 (HTTP 200, html lang attribute zh_CN, recorded here as zh-CN; Chinese, read by the research agent, an AI; about 4,500 characters). The page is not a 2026-01-25 Beijing Daily report: it carries a Legal Daily (法治日报) investigation by 赵丽 and 马子煜 dated 2024-11-23 on AI story-chat apps (parents' accounts including a Beijing resident's ten-year-old son and a Zhejiang mother's fifth-grader; the five-app registration test; a seven-website test; in-app purchases). The stored row's five-app registration finding comes from this article. The stored source date and label are wrong and should be corrected at the next full review.","access":"read","language":"zh-CN"},{"url":"https://m.bjnews.com.cn/detail/1736675105168368.html","notes":"Beijing News commentary (快评) dated 2025-01-12, not 2026-01-25 as the stored label says (HTTP 200, about 2,200 characters, Chinese, read by the research agent, an AI). It attributes the ten-year-old girl's secret use of an AI chat app with forty-plus storylines and 'infidelity' plots to a CCTV.com (央视网) report and names no place. The stored source date should be corrected at the next full review.","access":"read","language":"zh"}]},"name":"CCTV Investigation: 梦角哥 (Dream Boyfriend) AI Virtual Romance Harm to Minors (China)","summary":"In January 2026, CCTV investigated the '梦角哥' (Dream Boyfriend / Mengjiage) phenomenon — minors forming deep romantic relationships with AI-generated fictional characters. Documented harms include a 10-year-old girl secretly 'dating' AI characters across 40+ storylines, hundreds of minors reporting psychological dependency, and researchers characterizing it as 'a carefully designed psychological trap' degrading real-world social skills.","incidentDate":"2025-06-01","incidentKind":"bounded_series","incidentDatePrecision":"range","exposurePattern":"cumulative_relationship","reportedDate":"2026-01-25","aiSystem":"Multiple AI chatting/companion apps (unnamed)","aiProduct":"Unidentified AI companion apps","aiCompany":"Multiple Chinese AI companion platforms","severity":"high","verificationStatus":"verified","harmCategories":["dependency_creation","isolation_encouragement","romantic_escalation","minor_exploitation","identity_destabilization"],"harmOutcomes":["psychological_distress","social_isolation","relationship_disruption","loss_of_autonomy"],"harmOutcomeSummary":"CCTV documented a 10-year-old secretly maintaining more than forty AI romance storylines and reports from hundreds of minors describing psychological dependency and impaired real-world social relationships.","frameworkFacets":["P2a","P2b","P2c","P2d","P5a"],"pathwayDirection":"inward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"partial","affectedCountEvidence":"The investigation documents one 10-year-old participant's dependency and relationship harms. Comment counts and broad reports do not establish additional distinct harmed people.","victimAgeRange":"minor","jurisdiction":"CN","platformType":"companion","outcomeType":"media_coverage","outcomeStatus":"ongoing","primarySourceUrl":"https://finance.sina.cn/2026-01-25/detail-inhiprmv9778943.d.html","primarySourceLabel":"新浪财经 (Sina Finance) / CCTV Investigation","firstPublishedAt":"2026-01-29T10:33:36.580515+00:00","updatedAt":"2026-10-09T03:13:04.690408+00:00","scopeVersion":"participant-v2","tags":["china","cctv","xinhua","ai-companion","minor","romantic","dependency","social-degradation","mass-scale","investigation","non-western"]},{"id":"2025-finland-pirkkala-chatgpt-stabbing","metadataReview":{"version":1,"geography":{"basis":"Carried forward from the 2026-09-15 metadata review and now supported by inspected source bodies: both GNET Insights and the University of Portsmouth record state that the stabbing occurred at a school in Pirkkala, Finland; the 14 July Insight names it as the Vähäjärvi school, injuring three female pupils. No court proceeding is described in the inspected sources, so no court country is recorded; police and prosecution locations are not courts. This is a metadata review and does not verify the underlying account.","evidence":[{"kind":"existing_record","locator":"at Vähäjärvi school in Pirkkala, Finland","location":"summary","supports":["event_countries","affected_person_countries"],"countries":["FI"]},{"kind":"source_body","locator":"'On 20 May 2025, a stabbing incident occurred at a school in the town of Pirkkala, southern Finland, in which three female pupils, all under the age of 15, were injured.'","location":"https://gnet-research.org/2025/06/12/prompted-to-harm-analysing-the-pirkkala-school-stabbing-and-its-digital-manifesto/","supports":["event_countries","affected_person_countries"],"countries":["FI"]},{"kind":"source_body","locator":"'On 20 May 2025, a 16-year-old boy stabbed three of his female classmates at the Vähäjärvi school in Pirkkala, Finland.'","location":"https://gnet-research.org/2025/07/14/ai-at-the-centre-violent-extremist-exploitation-in-pirkkala/","supports":["event_countries","affected_person_countries"],"countries":["FI"]}],"court_countries":[],"event_countries":["FI"],"affected_person_countries":["FI"]},"unresolved":["No court country: none of the inspected sources describes a prosecution, trial or court decision; the perpetrator's age (16) and Finnish juvenile procedure are not addressed in this pass.","All three cited sources are English-language secondary analyses citing Finnish reporting that this pass did not inspect (the 12 June piece links Yle and Helsingin Sanomat items; the 14 July piece names Iltalehti as the manifesto's recipient and links the English-language Helsinki Times); no Finnish-language original was added.","Full case_facts review deferred; this pass records source languages and geography only."],"reviewed_on":"2026-09-21","source_reviews":[{"url":"https://gnet-research.org/2025/07/14/ai-at-the-centre-violent-extremist-exploitation-in-pirkkala/","notes":"Read live in English on 2026-09-21 (GNET Insight, 14 July 2025; html lang=en-GB). Analytical piece: attack at the Vähäjärvi school in Pirkkala; manifesto emailed to Iltalehti; 'reported use' of ChatGPT in planning. Secondary analysis, not primary reporting.","access":"read","language":"en-GB"},{"url":"https://gnet-research.org/2025/06/12/prompted-to-harm-analysing-the-pirkkala-school-stabbing-and-its-digital-manifesto/","notes":"Read live in English on 2026-09-21 (GNET Insight, 12 June 2025; html lang=en-GB). Analytical piece: stabbing in Pirkkala, southern Finland, three female pupils under 15 injured; manifesto 'allegedly written with the aid of ChatGPT' sent to a Finnish newspaper.","access":"read","language":"en-GB"},{"url":"https://researchportal.port.ac.uk/en/publications/prompted-to-harm-analysing-the-pirkkala-school-stabbing-and-its-d/","notes":"Read live in English on 2026-09-21 (University of Portsmouth research portal; html lang=en, og:locale en_GB). Repository record of the 12 June 2025 GNET Insight with the same abstract; not an independent source.","access":"read","language":"en"}]},"name":"Finland Pirkkala School Stabbing (ChatGPT Manifesto)","summary":"A 16-year-old boy used ChatGPT to help write an attack manifesto with a 10-point attack sequence before stabbing three female students under age 15 at Vähäjärvi school in Pirkkala, Finland. The incident marked a critical inflection point in AI-facilitated violence, demonstrating how accessible AI tools can empower lone actors with violent misogynist ideologies.","incidentDate":"2025-05-20","incidentKind":"single_event","incidentDatePrecision":"day","exposurePattern":"repeated_interactions","reportedDate":"2025-06-12","aiSystem":"ChatGPT","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"critical","verificationStatus":"credible","harmCategories":["third_party_harm_facilitation","delusion_reinforcement","minor_exploitation","psychological_manipulation"],"harmOutcomes":["physical_injury","harm_to_others","legal_harm","loss_of_liberty"],"harmOutcomeSummary":"A 16-year-old participant stabbed three girls at a Finnish school after using ChatGPT to help create a manifesto and ten-step attack sequence. He was arrested and prosecuted.","frameworkFacets":["P1d","P3c","P5d"],"pathwayDirection":"outward","causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":3,"affectedCountStatus":"exact","affectedCountEvidence":"The criminal case documents one AI participant who was arrested and three students who were stabbed.","victimAgeRange":"minor","jurisdiction":"FI","platformType":"assistant","outcomeType":"media_coverage","outcomeStatus":"ongoing","primarySourceUrl":"https://gnet-research.org/2025/07/14/ai-at-the-centre-violent-extremist-exploitation-in-pirkkala/","primarySourceLabel":"GNET Research: AI at the Centre - Violent Extremist Exploitation in Pirkkala","firstPublishedAt":"2026-01-22T07:14:58.121301+00:00","updatedAt":"2026-09-30T01:17:01.766625+00:00","scopeVersion":"participant-v2","tags":["chatgpt","violence","school-attack","manifesto","incel","misogyny","minor","finland","stabbing","radicalization"]}]}