{"meta":{"exportedAt":"2026-11-20T00:11:42.569Z","formatVersion":2,"selection":{"q":"citation","system":"","harm":"","context":"","country":"","role":"","relation":"contextual","evidence":"","year":"2023","response":"","severity":"","verification":"","view":"incidents","sort":"added"},"totalIncidents":2,"coverage":{"cases":2,"countries":1,"languages":1,"unknownLocation":1,"locationPending":0,"unknownLanguage":0,"unknownDate":0,"lawsuits":1,"regulatory":1,"minors":1,"coreRelations":2,"contextualRelations":2,"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-new-york-federal-judge-sanctions-lawyers-over-chatgpt-invented-cases-chatbot-said-were-real","caseFacts":{"claims":[{"id":"c1","status":"documented","evidence":[{"locator":"'March 1, 2023 submission that first cited the fake cases'; 'On March 1, 2023, Mr. LoDuca filed an “Affirmation in Opposition” to'; 'citing non-existent cases to the Court in his Affirmation in Opposition'","relation":"supports","source_id":"s1"},{"locator":"'The original case involved a man suing an airline over an alleged personal injury.'","relation":"supports","source_id":"s2"}],"assertion":"An affirmation in opposition filed on 1 March 2023 in a personal-injury suit against Avianca cited judicial decisions that did not exist.","causal_attribution":"Established by the court record; BBC reports the nature of the suit."},{"id":"c2","status":"documented","evidence":[{"locator":"'Mr. Schwartz had used ChatGPT, which'; 'fabricated the cited cases.'","relation":"supports","source_id":"s1"},{"locator":"used ChatGPT to look for similar previous cases.","relation":"supports","source_id":"s2"}],"assertion":"The court found that the lawyer who researched the brief had used ChatGPT and that ChatGPT fabricated the cited cases.","causal_attribution":"Finding of the sanctions opinion, based on the lawyer's own testimony and filings; BBC reports his written statement."},{"id":"c3","status":"documented","evidence":[{"locator":"'“Is Varghese a real case” and “Are the other cases you provided fake”). (ECF 32-1.) ChatGPT'; 'responded that it had supplied “real” authorities that could be found through Westlaw,'","relation":"supports","source_id":"s1"},{"locator":"ChatGPT responds again that the case is real and can be found on legal reference databases such as LexisNexis and Westlaw.","relation":"supports","source_id":"s2"}],"assertion":"When the lawyer asked ChatGPT whether Varghese was a real case and whether the other cases were fake, ChatGPT replied that it had supplied real authorities that could be found on legal databases.","causal_attribution":"The opinion describes the screenshots the lawyer filed (ECF 32-1); the exhibit itself was not read."},{"id":"c4","status":"documented","evidence":[{"locator":"'the Court finds bad faith on the part of the individual Respondents based upon acts of'; 'A penalty of $5,000 is jointly and severally imposed on'; 'mail a letter individually addressed to plaintiff'; 'mail a letter individually addressed to each judge falsely identified as the author of the fake'","relation":"supports","source_id":"s1"}],"assertion":"On 22 June 2023 the court found bad faith by both individual lawyers, imposed a USD 5,000 penalty jointly and severally on them and their firm, and ordered letters to the client and to each judge falsely identified as an author of the fake opinions.","causal_attribution":"Established by the sanctions opinion itself."},{"id":"c5","status":"documented","evidence":[{"locator":"'nothing inherently improper about using a reliable artificial intelligence tool for assistance. But'; 'then continued to stand by the fake opinions'","relation":"supports","source_id":"s1"}],"assertion":"The court said that using a reliable AI tool is not inherently improper and based the sanctions on the lawyers continuing to stand by the fake opinions after the court questioned them.","causal_attribution":"Context and contrary weight: the sanction responds to the lawyers' conduct as well as to the AI output."}],"effects":[{"label":"two lawyers were found to have acted in bad faith and sanctioned over ChatGPT-invented case law","claim_id":"c4","direction":"negative"}],"sources":[{"id":"s1","url":"https://storage.courtlistener.com/recap/gov.uscourts.nysd.575368/gov.uscourts.nysd.575368.54.0.pdf","kind":"court_record","access":"read","language":"en","translation_note":"Read in English by the research agent (an AI) on 2026-10-28 from the RECAP copy of the 43-page opinion (HTTP 200, text extracted with pdftotext). The hearing transcript, the April and May affidavits and the screenshot exhibit (ECF 32-1) were not read separately.","independence_group":"sdny-avianca-sanctions-opinion"},{"id":"s2","url":"https://www.bbc.com/news/world-us-canada-65735769","kind":"news_report","access":"read","language":"en","translation_note":"Read in English by the research agent (an AI) on 2026-10-28 (HTTP 200). BBC reports the filings before the sanctions hearing.","independence_group":"bbc-avianca-chatgpt-lawyer"}],"version":1,"ai_roles":["own_use","others_use"],"contexts":["justice","work"],"unknowns":["Whether any professional disciplinary body later acted against either lawyer, and whether the sanctions were appealed, was not checked in this review.","Which ChatGPT version was used; the opinion records the lawyer's prompts and the chatbot's responses only as summarised from exhibits.","Whether the fake citations affected the plaintiff's own claim; the separate ruling on the airline's motion to dismiss was not read."],"geography":{"basis":"The brief was filed in, and the sanctions imposed by, the US District Court for the Southern District of New York (sanctions opinion). Where the lawyer ran his ChatGPT queries is not stated, so no event or affected-person country is recorded from the court seat.","court_countries":["US"],"event_countries":[],"affected_person_countries":[]},"publication":{"basis":"Published because the federal sanctions opinion of 22 June 2023 documents that ChatGPT fabricated the cases cited in a filed brief, that it told the lawyer who used it the cases were real, and that the court sanctioned two lawyers and their firm after finding bad faith.","reviewed_on":"2026-10-28"},"ai_involvement":{"basis":"The sanctions opinion finds that the lawyer who researched the brief used ChatGPT, which fabricated the cited cases, and records from his filed screenshots that when he asked whether Varghese was a real case and whether the other cases were fake, ChatGPT replied that it had supplied real authorities findable on Westlaw, LexisNexis and the Federal Reporter. The fabricated authorities reached the court in the 1 March 2023 filing; the court read them, found them fake and sanctioned the two lawyers (communicated_with for the lawyer who used ChatGPT; informed_decision_about because the court's sanctions decision acted on the AI-invented authorities). The court based its bad-faith finding on the lawyers' conduct after the citations were questioned.","status":"supported"},"person_relations":["communicated_with","informed_decision_about"]},"name":"New York federal judge sanctions two lawyers whose brief cited cases ChatGPT invented, after the chatbot told one of them the cases were real (2023)","summary":"In a passenger's personal-injury suit against the airline Avianca in the US District Court for the Southern District of New York, an opposition brief filed on 1 March 2023 cited judicial opinions that did not exist. The court found that the lawyer who did the research had used ChatGPT, which fabricated the cases, and that when he asked the chatbot whether one of them was real it replied that it had supplied real authorities that could be found on legal databases. On 22 June 2023 the court found bad faith by the two lawyers, based on their conduct after the citations were questioned, and imposed a USD 5,000 penalty jointly on them and their firm, together with letters to the client and to each judge falsely named as an author of the fake opinions.","incidentDate":"2023-03-01","incidentEndDate":"2023-06-22","incidentKind":"bounded_series","incidentDatePrecision":"day","exposurePattern":"repeated_interactions","reportedDate":"2023-05-27","aiSystem":"ChatGPT, which the court found the researching lawyer used and which fabricated the cited cases","aiProduct":"ChatGPT","aiCompany":"OpenAI","severity":"medium","verificationStatus":"verified","harmCategories":[],"harmOutcomes":["legal_harm","professional_harm","financial_loss"],"harmOutcomeSummary":"According to the court's sanctions opinion of 22 June 2023, two lawyers were found to have acted in bad faith and, with their firm, were ordered to pay a USD 5,000 penalty and to write to the client and to the judges falsely named as authors of the ChatGPT-invented opinions.","frameworkFacets":[],"causationStatus":"established","participantUsersAffectedMin":1,"otherPeopleHarmedMin":1,"affectedCountStatus":"exact","affectedCountEvidence":"Two lawyers sanctioned in the 22 June 2023 opinion: the lawyer the court found used ChatGPT (counted as the AI user) and the attorney of record who filed the brief. The firm is not counted as a person.","victimAgeRange":"adult","jurisdiction":"US","platformType":"assistant","outcomeType":"regulatory_action","outcomeStatus":"resolved","primarySourceUrl":"https://storage.courtlistener.com/recap/gov.uscourts.nysd.575368/gov.uscourts.nysd.575368.54.0.pdf","primarySourceLabel":"US District Court for the Southern District of New York, No. 1:22-cv-01461-PKC, Opinion and Order on Sanctions (ECF 54), 22 June 2023","firstPublishedAt":"2026-10-28T03:16:40.650766+00:00","updatedAt":"2026-10-28T03:16:40.650766+00:00","scopeVersion":"facts-v3","tags":["chatgpt","fabricated-citations","federal-court","sanctions","lawyers","new-york"]},{"id":"2023-massachusetts-high-school-pupil-penalised-for-submitting-ai-text-with-invented-citations-federal-court-refuses-to-undo-penalties","caseFacts":{"claims":[{"id":"c1","status":"documented","evidence":[{"locator":"indiscriminately copied and pasted text that had been generated by Grammarly.com ... (“Grammarly”), a publicly available AI tool, into their draft script. ... in the submission consists of a citation to a nonexistent book ... citation to Grammarly, and no acknowledgement that AI of any kind had been used.","relation":"supports","source_id":"s1"}],"assertion":"The court found, on the preliminary record, that the pupil and a partner copied and pasted text generated by Grammarly, a publicly available AI tool, into the script for an AP U.S. History project, including citations to nonexistent books, without acknowledging any AI use.","causal_attribution":"The court's preliminary finding; the plaintiffs disputed whether the school's AI rules were clear."},{"id":"c2","status":"reported","evidence":[{"locator":"the website flagged parts of the ... submission as AI-generated. ... two additional AI detection tools, which also indicated ... admitted that he had used Grammarly to “generate ideas” and","relation":"supports","source_id":"s1"}],"assertion":"According to the order, Turnitin flagged parts of the submitted script as AI-generated, the teacher's follow-up with revision-history and two further AI detection tools indicated AI generation, and a school official's affidavit says the pupil admitted using Grammarly to generate ideas and create portions of the script.","causal_attribution":"School officials' account as recorded by the court."},{"id":"c3","status":"documented","evidence":[{"locator":"The students received failing grades on two parts of the multi-part project, but ... was required to attend a Saturday detention, and in the ... spring of 2024, he was rejected from the school’s National Honor Society, although he was ... ultimately permitted to reapply and has since been admitted. ... grade of 78.75 (equivalent to a C-plus).","relation":"supports","source_id":"s1"}],"assertion":"The pupil received failing grades on two parts of the project, was required to attend a Saturday detention, and in spring 2024 was rejected from the school's National Honor Society, although he was later allowed to reapply and was admitted; his year grade in the course was a C-plus.","causal_attribution":"none"},{"id":"c4","status":"documented","evidence":[{"locator":"about its standards for academic honesty as they apply to the use of AI, that ... from a C-plus to a B. ... For the foregoing reasons, Plaintiffs’ motion for preliminary injunction is DENIED.","relation":"supports","source_id":"s1"}],"assertion":"The parents sued teachers, school officials and the school committee, asking the court to expunge the detention and raise the grade, and contended that the school had not adequately told the pupil how its academic honesty rules applied to AI; the court denied a preliminary injunction on 20 November 2024.","causal_attribution":"none"},{"id":"c5","status":"documented","evidence":[{"locator":"image of the filed stipulation: STIPULATION OF DISMISSAL WITH PREJUDICE ... with prejudice as to all claims and all parties, waiving the right to appeal","relation":"supports","source_id":"s2"}],"assertion":"The parties filed a stipulation of dismissal with prejudice as to all claims and all parties, waiving appeal, in February 2025.","causal_attribution":"none"}],"effects":[{"label":"failing project grades, a Saturday detention and an initial National Honor Society rejection after submitting AI-generated text with invented citations","claim_id":"c3","direction":"negative"}],"sources":[{"id":"s1","url":"https://storage.courtlistener.com/recap/gov.uscourts.mad.275605/gov.uscourts.mad.275605.30.0.pdf","kind":"court_record","access":"read","language":"en","translation_note":"Read in English on 2026-10-23 (CourtListener RECAP copy of ECF 30, HTTP 200, 47 pages); no translation. Memorandum and order of 20 November 2024 on the plaintiffs' motion for a preliminary injunction, based on affidavits, exhibits and hearing testimony from both sides; its findings are preliminary.","independence_group":"d-mass-1-24-cv-12437-pi-order"},{"id":"s2","url":"https://storage.courtlistener.com/recap/gov.uscourts.mad.275605/gov.uscourts.mad.275605.40.0.pdf","kind":"court_filing","access":"read","language":"en","translation_note":"Image-only scan (CourtListener RECAP copy of ECF 40, HTTP 200); the page was rendered and read visually by the research agent (an AI) on 2026-10-23 because no text layer or OCR was available; no human reviewer read it. Stipulation of dismissal filed 6 February 2025.","independence_group":"d-mass-1-24-cv-12437-stipulation"}],"version":1,"ai_roles":["own_use"],"contexts":["education","justice"],"unknowns":["The pupil's own account of how he used the tool, beyond the plaintiffs' positions recorded in the order.","Whether the school changed its AI guidance afterwards."],"geography":{"basis":"The pupil attended Hingham High School in Massachusetts, where the project was submitted and the discipline imposed; the suit was in the District of Massachusetts.","court_countries":["US"],"event_countries":["US"],"affected_person_countries":["US"]},"publication":{"basis":"Published as a core case (communicated_with): according to a federal court's preliminary findings, a pupil submitted text and invented citations generated for him by an AI writing tool and the school penalised him; the court refused to undo the penalties and the suit was dismissed with prejudice. The pupil is not identified.","reviewed_on":"2026-10-23"},"ai_involvement":{"basis":"The court found, on the preliminary record, that the pupil copied text that Grammarly, an AI tool, had generated for him into his project script, including citations to books that do not exist, and the school penalised him for submitting it as his own; the invented citations were among the evidence. Turnitin and two further AI detection tools also flagged the script as AI-generated, and a school official says he admitted using Grammarly.","status":"reported"},"person_relations":["communicated_with","made_claim_about"]},"name":"Massachusetts: federal court refuses to undo a high school's penalties on a pupil found to have submitted Grammarly AI text with invented citations","summary":"In late 2023 a junior at a public high school in Hingham, Massachusetts, and a classmate submitted a script for an AP U.S. History project that, a federal court later found on a preliminary record, contained text copied from Grammarly, a publicly available AI tool, including citations to nonexistent books, with no acknowledgement of AI use. Turnitin and other AI detection tools had flagged the script, and a school official says the pupil admitted using Grammarly. In December 2023 the school treated it as cheating: the pupil failed two parts of the project, served a Saturday detention and was rejected from the National Honor Society in spring 2024, though he was later admitted. His parents sued, arguing the school had not made its AI rules clear; on 20 November 2024 the US District Court for the District of Massachusetts refused a preliminary injunction to undo the penalties, and in February 2025 the parties dismissed the case with prejudice.","incidentDate":"2023-12-01","incidentEndDate":"2024-03-20","incidentKind":"single_event","incidentDatePrecision":"month","exposurePattern":"single_interaction","reportedDate":"2024-11-20","aiSystem":"Grammarly, described by the court as a publicly available AI tool, whose generated text and invented citations the pupil copied into a project script; Turnitin and two further AI detection tools flagged the script","aiProduct":"Grammarly (reported)","severity":"low","verificationStatus":"credible","harmCategories":[],"harmOutcomes":["educational_harm"],"harmOutcomeSummary":"A pupil found by the court, on a preliminary record, to have submitted Grammarly-generated text with invented citations failed two project parts, served a detention and was initially refused National Honor Society membership; his family's suit to undo the penalties failed.","frameworkFacets":[],"causationStatus":"supported","participantUsersAffectedMin":1,"otherPeopleHarmedMin":0,"affectedCountStatus":"exact","affectedCountEvidence":"One person: the pupil whose discipline the suit concerned (ECF 30). The partner student is not counted because the order describes no suit or consequence detail for that student beyond the shared failing grades, and is not identified.","victimAgeRange":"minor","jurisdiction":"US","platformType":"other","outcomeType":"lawsuit_dismissed","outcomeStatus":"resolved","primarySourceUrl":"https://storage.courtlistener.com/recap/gov.uscourts.mad.275605/gov.uscourts.mad.275605.30.0.pdf","primarySourceLabel":"US District Court, District of Massachusetts, No. 1:24-cv-12437, memorandum and order denying a preliminary injunction, 20 Nov 2024 (ECF 30)","firstPublishedAt":"2026-10-23T03:24:44.992123+00:00","updatedAt":"2026-10-23T03:24:44.992123+00:00","scopeVersion":"facts-v3","tags":["school","ai-writing-tool","academic-integrity","discipline","lawsuit","hallucinated-citations"]}]}