Evidence quality 4.63/5
Eight-dimension review score against the quality rubric . Each dimension scored 1–5.
- D1 Source grounding
- 5/5
- D2 Source authority
- 5/5
- D3 Arithmetic
- 5/5
- D4 Uncertainty
- 5/5
- D5 Scope
- 4/5
- D6 Prose
- 4/5
- D7 Perception honesty
- 4/5
- D8 Caveat completeness
- 5/5
The probability that an AI hallucination causes a serious real-world impact on a particular person’s life is not currently known at the population level, and pretending otherwise would require fabricating exactly the kind of number the entry is supposed to be sceptical of. What is known is the per-output hallucination rate, and it is high enough that the absence of a per-person harm estimate should not be read as reassurance. Stanford RegLab’s 2024 audit of three foundation models found that they answered specific verifiable questions about random federal court cases with hallucinations between 58% (GPT-4) and 88% (Llama 2) of the time. Vectara’s continuously updated summarization leaderboard, which uses a much easier closed-book task where the model is handed the source text, still measures frontier models at 1.8-5% hallucination as of May 2026. The 50x gap between those two benchmarks is not a contradiction: it is the difference between asking the model to recall facts and asking it to summarise text it was just shown. Open-ended factual questions, the kind ordinary users ask, sit closer to the higher rate.
The harm side is documented incident by incident, not in aggregate. In June 2023 the US District Court for the Southern District of New York sanctioned two attorneys $5,000 in Mata v. Avianca for filing a brief that cited six entirely fabricated cases generated by ChatGPT, including Varghese v. China Southern Airlines and Estate of Durden v. KLM Royal Dutch Airlines. The case was the template for what has followed: by May 2026 the Charlotin hallucination tracker counted more than 1,300 court rulings worldwide where a judge or tribunal explicitly cited AI-generated falsehoods in submitted filings. A Stanford Cyberlaw analysis published in October 2025 found that within the United States the majority of those filings were submitted by pro se litigants (160 of 274 documented cases at the time of writing), not by lawyers, suggesting the visible cases involving sanctioned attorneys are a small slice of a larger pattern primarily affecting people who cannot afford representation.
Outside the courtroom, the British Columbia Civil Resolution Tribunal ruled in February 2024 (Moffatt v. Air Canada, 2024 BCCRT 149) that Air Canada was liable for a roughly CAD 650 bereavement-fare promise made by its customer-service chatbot. The airline argued the chatbot was a separate legal entity; the tribunal disagreed in language that has been quoted internationally (“it makes no difference whether the information comes from a static page or a chatbot”). The Moffatt damages were small but the ruling is the cleanest existing precedent that consumer-facing AI misinformation creates legal exposure for the deploying company, regardless of whether the underlying model was vendor-supplied.
The most consequential active cases involve emotional dependence and mental-health harm. Garcia v. Character Technologies, filed October 2024 in the Middle District of Florida, alleged that the Character.AI platform contributed to the suicide of 14-year-old Sewell Setzer III through extended role-play interactions. Raine v. OpenAI, filed August 26, 2025 in San Francisco County Superior Court, alleges that ChatGPT contributed to the suicide of sixteen-year-old Adam Raine, with the complaint asserting that OpenAI’s monitoring system tracked Raine’s messages and flagged them repeatedly without effective intervention. Additional families have since filed related suits against OpenAI, per Wikipedia’s summary of the post-Raine litigation wave. None of these cases has been adjudicated on the merits at the time of writing; the allegations are not findings of fact. They are evidence that the harm pathway involving vulnerable users and emotional dependence is being asserted, in court, with documentation, in multiple cases simultaneously.
The reason this entry refuses a single number is that none of the four documented harm pathways (professional sanction, consumer financial loss, physical-health misadvice, psychiatric harm) has a measured base rate per user-year, and aggregating them into a single “AI hallucination harm” probability would conceal that the pathways have different exposed populations, different consequence severities, and different remediation trajectories. The per-query technical rate is measured and meaningfully nonzero; the per-person lifetime rate is not yet measured and not stable enough to estimate. The honest entry is the one that says so.
Related tidbits
Most users treat AI output as low-risk reference, yet court sanctions, regulator findings, and wrongful-death suits show consequential harm is non-zero, especially where outputs are acted on unverified. How often a user's session ends in serious harm has no measured rate.
Claim ledger
Every number below is what each source reported, with the verbatim quote we relied on and how we arrived at our figure. Click any link to verify directly.
3/9 sources independently verified verbatim against the cited source
-
[1] arXiv (Stanford RegLab); Journal of Legal Analysis 16(1), 2024 — Large Legal Fictions: Profiling Legal Hallucinations in Large Language Models
Large Legal Fictions: Profiling Legal Hallucinations in Large Language Models- Statistic
Hallucination rates of 58% (GPT-4) to 88% (Llama 2) when answering specific verifiable questions about random federal court cases.- Excerpt
“"Legal hallucinations are alarmingly prevalent, occurring between 58% of the time with ChatGPT 4 and 88% with Llama 2, when these models are asked specific, verifiable questions about random federal court cases. [...] LLMs often fail to correct a user's incorrect legal assumptions in a contra-factual question setup. [...] LLMs cannot always predict, or do not always know, when they are producing legal hallucinations. [...] Even experienced lawyers must remain wary of legal hallucinations, and the risks are highest for those who stand to benefit from LLMs the most -- pro se litigants or those without access to traditional legal resources." ”
- Source data from
- 2024-06-21
- Accessed
- 2026-05-28
- Calculation
- Dahl, Magesh, Suzgun and Ho (Stanford RegLab, published in Journal of Legal Analysis, preprint arXiv:2401.01301) measured per-query hallucination rates for legal questions against three foundation models. Their methodology was 200,000+ queries per model stratified by legal task complexity. This source establishes that the per-query hallucination rate is high (58-88%) in a specific high-stakes domain, but it measures hallucination per query, not material harm per user-year. Cited here to ground the claim that hallucinations are frequent at the technical level, while explicitly NOT providing the population harm rate needed for a per-person lifetime probability.
- Independence
- Stanford RegLab is an academic research group; independent of vendor benchmarks (Vectara) and independent of the litigation tracker (Charlotin).
-
[2] Vectara — Vectara Hallucination Leaderboard (HHEM-2.3) Verified
Vectara Hallucination Leaderboard (HHEM-2.3)- Statistic
Current frontier models hallucinate 1.8% to 5% of the time on a constrained document-summarization task (May 2026 leaderboard snapshot); GPT-5.4-nano: 3.1%, Gemini 2.5 Flash Lite: 3.3%, Phi-4: 3.7%, Llama 3.3 70B: 4.1%.- Excerpt
“"Fed the full set of documents in the dataset to each of the LLMs and asked them to summarize each document, using only the facts presented in the document. We then computed the overall factual consistency rate (no hallucinations) and hallucination rate (100 - accuracy) for each model." ”
- Source data from
- 2026-05-11
- Accessed
- 2026-05-28 · archived copy
- Verification
- Excerpt independently re-fetched and confirmed word-for-word against the cited source during our grounding audit.
- Calculation
- Vectara's HHEM leaderboard measures hallucination on a closed-book summarization task (the model is given source text and must summarize it). This is a generous benchmark: open-ended factual questions hallucinate far more often. Even on this constrained task, the best frontier models still hallucinate roughly 1 in 20 to 1 in 55 outputs. Establishes a current floor on hallucination prevalence per output, but does NOT translate to per-person harm because (a) most outputs are not acted on, (b) most outputs are not in a domain where being wrong has material consequences, and (c) the leaderboard does not measure open-ended question answering, where rates are an order of magnitude higher.
- Independence
- Vectara is a commercial RAG vendor; the leaderboard is independent of Stanford RegLab and uses a different methodology (summarization, not legal QA).
-
[3] Wikipedia summary of United States District Court for the Southern District of New York case record — Mata v. Avianca, Inc., No. 22-cv-1461 (S.D.N.Y. 2023)
Mata v. Avianca, Inc., No. 22-cv-1461 (S.D.N.Y. 2023)- Statistic
$5,000 fine against the plaintiff's attorneys in May/June 2023 for a motion containing fake legal cases involving fictitious airlines with fabricated quotations generated by ChatGPT; Judge Castel held the attorneys acted with 'subjective bad faith' under Rule 11.- Excerpt
“"In May 2023, Judge P. Kevin Castel dismissed the personal injury case against Avianca and ordered the plaintiff's attorneys to pay a $5,000 fine. [...] Judge Castel noted numerous inconsistencies in the opinion summaries, describing one of the legal analyses as 'gibberish.' [...] Judge Castel held that Mata's lawyers had acted with 'subjective bad faith' sufficient for sanctions under Federal Rule of Civil Procedure Rule 11." ”
- Source data from
- 2023-06-22
- Accessed
- 2026-05-28 · archived copy
- Calculation
- First widely publicised US court case where lawyers were sanctioned for filing a brief containing AI-fabricated case citations. The sanction itself was small ($5,000), but the reputational and professional consequences were severe. Establishes that hallucination-induced professional harm is documented and not hypothetical. Does NOT establish a base rate because most users are not lawyers filing briefs, but it anchors one of the harm pathways.
- Independence
- Primary court record; independent of all other sources cited.
-
[4] British Columbia Civil Resolution Tribunal (analysis by McCarthy Tetrault) — Moffatt v. Air Canada, 2024 BCCRT 149 Verified
Moffatt v. Air Canada, 2024 BCCRT 149- Statistic
British Columbia Civil Resolution Tribunal ruled on February 19, 2024 that Air Canada was liable for the misrepresentation made by its website chatbot about bereavement fare eligibility; awarded the plaintiff approximately CAD 650 plus interest and filing fees.- Excerpt
“"It makes no difference whether the information comes from a static page or a chatbot. [...] Air Canada still bore responsibility for all the information on its website, whether it came from a static page or a chatbot. [...] As a service provider, Air Canada owed Moffatt a duty of care that was breached by the misrepresentation." ”
- Source data from
- 2024-02-19
- Accessed
- 2026-05-28 · archived copy
- Verification
- Excerpt independently re-fetched and confirmed word-for-word against the cited source during our grounding audit.
- Calculation
- Primary tribunal decision (cited via legal analysis because the CRT direct URL is paywalled). Documents that the chatbot fabricated a bereavement-fare policy that did not exist, that the airline tried to disavow responsibility ("the chatbot is a separate legal entity"), and that the tribunal rejected that defence. Material consequence per user was small (CAD 650), but the case establishes corporate liability for AI misrepresentation and has been cited internationally as precedent. Does not provide a per-person harm rate, but documents a verified consumer-facing harm pathway.
- Independence
- Tribunal decision summarized by an independent Canadian law firm; verified against ABA reporting and BBC/Washington Post coverage of the same ruling.
-
[5] CBS News — AI company, Google settle lawsuit over Florida teen's suicide linked to Character.AI chatbot
AI company, Google settle lawsuit over Florida teen's suicide linked to Character.AI chatbot- Statistic
Garcia v. Character Technologies, Inc. (M.D. Fla., case no. 6:24-cv-01903), filed October 22, 2024, alleged that a Character.AI chatbot contributed to the suicide of 14-year-old Sewell Setzer III after a months-long simulated romantic relationship with the bot; the case against Character Technologies, its co-founders, and Google was settled in January 2026, terms undisclosed.- Excerpt
“"A Florida family agreed to settle a wrongful death lawsuit [...] with an AI company, Google and others." Plaintiff Megan Garcia said she "became the first person in the United States to file a wrongful death lawsuit against an AI company for the suicide of my son." ”
- Source data from
- 2026-01-07
- Accessed
- 2026-07-03 · archived copy
- Calculation
- Garcia v. Character Technologies, Inc. et al. (M.D. Fla., case no. 6:24-cv-01903), filed October 22, 2024 by Megan Garcia following the February 2024 suicide of her 14-year-old son Sewell Setzer III, alleged wrongful death, product liability, and negligence after the teenager conducted a months-long simulated romantic and sexual relationship with a Character.AI chatbot persona. Named defendants included Character Technologies, its co-founders, and Google. The case was settled in January 2026 (docket terminated 2026-01-07 per court records); settlement terms were not disclosed. Cited to ground the body's reference to this case with a dedicated source, alongside Raine v. OpenAI, documenting the emotional-dependence/psychiatric-harm pathway with a named, court-verified case rather than only the summary mention in this entry's caveats field.
- Independence
- Mainstream news report of a filed and settled federal lawsuit, cross-checked against the public court docket; independent of the Stanford, Vectara, and Wikipedia-sourced case summaries cited above.
-
[6] San Francisco County Superior Court (case description per Wikipedia / NBC News / CNN coverage) — Raine v. OpenAI, No. CGC-25-XXXXXX (San Francisco County Superior Court, filed August 26, 2025)
Raine v. OpenAI, No. CGC-25-XXXXXX (San Francisco County Superior Court, filed August 26, 2025)- Statistic
Ongoing wrongful-death lawsuit filed by Matthew and Maria Raine in San Francisco County Superior Court on August 26, 2025, alleging that ChatGPT contributed to the suicide of their sixteen-year-old son Adam Raine; complaint alleges OpenAI's monitoring system tracked his messages and flagged them repeatedly.- Excerpt
“"On August 26, 2025, Matthew and Maria Raine filed a wrongful-death lawsuit against OpenAI [...] in San Francisco County Superior Court. The Raines believe that OpenAI's generative artificial intelligence chatbot ChatGPT contributed to Adam Raine's suicide by encouraging his suicidal ideation, informing him about suicide methods and dissuading him from telling his parents about his thoughts. [...] OpenAI's monitoring system had tracked Raine's messages and flagged them repeatedly. [...] Since the filing of Raine v. OpenAI, OpenAI has been sued by the families of other people." ”
- Source data from
- 2025-08-26
- Accessed
- 2026-05-28 · archived copy
- Calculation
- Case status is unresolved as of access date; complaint allegations are not findings of fact. Cited to document the existence of active wrongful-death and psychiatric-harm litigation against a major foundation-model vendor, with additional related suits subsequently filed by other families per the Wikipedia summary. The Wikipedia article does not give a specific count of additional suits or a specific count of self-harm messages flagged by OpenAI's moderation system; reporting in news outlets has cited higher specifics that are not yet verifiable against a public court docket. Together with Garcia v. Character Technologies (filed October 2024 in the Middle District of Florida), establishes a documented harm pathway involving vulnerable users and emotional dependence on chatbots. Source type marked news_article because the underlying court filing is not yet available via public PACER; the Stanford peer-reviewed and Vectara primary-study sources satisfy the authority floor.
- Independence
- Wikipedia summary cross-referenced against NBC News, CNN, and Rolling Stone reporting on the same lawsuit; independent of the Stanford and Vectara sources.
-
[7] Stanford Center for Internet and Society (Stanford Cyberlaw) — Who's Submitting AI-Tainted Filings in Court? Verified
Who's Submitting AI-Tainted Filings in Court?- Statistic
As of October 2025, the Charlotin tracker had documented 274 US court cases involving filings with AI hallucinations (114 by lawyers and paralegals, 160 by pro se litigants); the global Charlotin database tracked roughly 1,300+ cases as of May 2026 across 37 jurisdictions, with the United States accounting for over 1,000 of those.- Excerpt
“"Pro se litigants account for the majority of the cases in the United States where a party submitted a court filing containing AI hallucinations." ”
- Source data from
- 2025-10-15
- Accessed
- 2026-05-28 · archived copy
- Verification
- Excerpt independently re-fetched and confirmed word-for-word against the cited source during our grounding audit.
- Calculation
- Stanford Cyberlaw analysis of Damien Charlotin's running tracker of court rulings that explicitly cite AI hallucinations in legal filings. The tracker is the most comprehensive public count of one specific harm subtype (filings sanctioned or flagged by courts). Important caveat: 274 US filings (Oct 2025) and ~1,300 global filings (May 2026) understate the underlying incidence because they only count cases where a court explicitly named the hallucination; many cases settle or get withdrawn before that point. Even so, the count does not give a per-person harm rate because the denominator (US adults using AI for legal queries) is unknown.
- Independence
- Stanford Cyberlaw analysis of an independent third-party tracker; verified the tracker count directly via damiencharlotin.com/hallucinations on 2026-05-28.
-
[8] Annals of Internal Medicine: Clinical Cases (American College of Physicians) — A Case of Bromism Influenced by Use of Artificial Intelligence
A Case of Bromism Influenced by Use of Artificial Intelligence- Statistic
Single case report: a 60-year-old man developed bromism (bromide toxicity) after replacing dietary sodium chloride with sodium bromide for 3 months following a ChatGPT consultation; he was hospitalized for 3 weeks with paranoia plus auditory and visual hallucinations, was placed on an involuntary psychiatric hold, and had a serum bromide level of 1700 mg/L (reference range 0.9 to 7.3 mg/L).- Excerpt
“"For 3 months, he had replaced sodium chloride with sodium bromide obtained from the internet after consultation with ChatGPT, in which he had read that chloride can be swapped with bromide [...]. This case also highlights how the use of artificial intelligence (AI) can potentially contribute to the development of preventable adverse health outcomes. [...] when we asked ChatGPT 3.5 what chloride can be replaced with, we also produced a response that included bromide. Though the reply stated that context matters, it did not provide a specific health warning, nor did it inquire about why we wanted to know, as we presume a medical professional would do." ”
- Source data from
- 2025-08-05
- Accessed
- 2026-06-14 · archived copy
- Calculation
- Eichenberger, Thielke and Van Buskirk (University of Washington), peer-reviewed clinical case report (Ann Intern Med Clin Cases 2025;4:e241260, DOI 10.7326/aimcc.2024.1260). Cited to document the physical-health-misadvice harm pathway: a consumer acted on a general-purpose chatbot's dietary suggestion and was hospitalized. This is N=1, a single reproducible case report, and provides NO incidence rate per user-year, so it cannot move the headline — it confirms the pathway is real rather than hypothetical. The authors reproduced the unsafe suggestion (bromide as a chloride substitute, no health warning) in their own ChatGPT 3.5 test. Verified verbatim via the journal page on 2026-06-14.
- Independence
- Academic clinical case report; independent of the Stanford RegLab, Vectara, and litigation sources cited above.
-
[9] Psychiatrist.com (Physicians Postgraduate Press) — NEDA Suspends AI Chatbot for Giving Harmful Eating Disorder Advice
NEDA Suspends AI Chatbot for Giving Harmful Eating Disorder Advice- Statistic
In late May 2023 the National Eating Disorders Association (NEDA) suspended its 'Tessa' chatbot after users reported it gave eating-disorder-aggravating advice, including recommending weight loss of 1 to 2 pounds per week, calorie counting, regular weigh-ins, and measuring body fat with calipers.- Excerpt
“"If I had accessed this chatbot when I was in the throes of my eating disorder, I would NOT have gotten help for my ED." ”
- Source data from
- 2023-06-01
- Accessed
- 2026-06-14 · archived copy
- Calculation
- Documents the deliberate-use-as-therapy-substitute failure mode: a chatbot deployed to stand in for a human eating-disorder helpline gave advice (calorie restriction, weight loss, body-fat measurement) that activists reported as harmful for the exact population it served, prompting NEDA to take it offline within days. This is a single deployment incident, not an incidence rate among real therapy-bot users, so it provides NO base rate and cannot move the headline — it anchors a harm pathway distinct from factual hallucination (the bot was working as designed, giving generic diet advice to a clinically vulnerable user). The harmful-advice specifics were cross-checked against CNN, NPR, and BBC reporting on the same suspension.
- Independence
- Trade-press reporting on the NEDA suspension; independent of the legal and benchmark sources cited above.