Case files.
One outcome per file, checked, not claimed. Four kinds of file live here: deployments with figures, post-mortems of walk-backs, myth checks of famous claims, and first looks at new methods. Every figure is validated against the public record and the method ships on the page.
AlphaFold: independently ranked the top method in the CASP14 blind assessment (summed z-score 244 vs 91 for second place), scaled to 214M+ predicted structures, and won the 2024 Nobel Prize in Chemistry
In CASP14 (2020), a blind protein-structure-prediction assessment run by independent academic organizers, Google DeepMind's AlphaFold was ranked first by a wide margin — a summed z-score of 244.02 against 90.82 for the second-placed group — and reported a median score of 92.4 GDT across all targets.
Bank of Korea (Issue Note 2026-12): AI adoption cut work time 3.8% but the productivity gain is near zero
In a June 2026 Issue Note the Bank of Korea found that generative-AI adoption reduced average work time by 3.8 percent, about 1.5 hours per week, yet the relationship between those time savings and actual output growth was essentially zero, implying a potential productivity gain of only about 1.0 percent.
Court of Rome annuls the Garante's €15M ChatGPT fine against OpenAI on one-stop-shop jurisdiction (R.G. 4785/2025, 2026)
In November 2024 Italy's data-protection authority fined OpenAI €15,000,000 over ChatGPT and ordered a first-of-its-kind six-month public information campaign; on 18 March 2026 the Court of Rome annulled the whole decision, ruling the Garante had lost jurisdiction under the GDPR one-stop-shop once OpenAI's Irish establishment became its lead authority before the final decision.
The AI shopping app that ran on people: DOJ and SEC charge Nate's founder over an 'effectively zero percent' automation rate
On 9 April 2025 the U.S.
The 'AI-powered' store that lost its investors everything: the FTC banned Automators AI and entered a $21.7M judgment
The FTC sued Automators AI, a business-opportunity scheme that sold 'done-for-you' Amazon and Walmart storefronts marketed as powered by proprietary 'AI machine learning'.
GEMA v Suno (LG München I, 42 O 763/25): Munich court rules AI music training and model memorisation infringe copyright
On 31 July 2026 the Landgericht München I largely upheld GEMA's claims for cessation, disclosure and damages against the AI music generator Suno, holding that six protected works were reproduced during training and memorised in Suno's models v3.5 and v4, with further infringement through generated outputs, a judgment that is not yet final.
A German court held Google directly liable for its AI Overviews: the Landgericht München I ruled the AI summary is Google's own speech, not a search result (26 O 869/26)
On 28 May 2026 the Landgericht München I issued a preliminary injunction finding Google directly liable for defamatory statements its AI Overviews feature generated about two Munich publishers, holding that an AI-generated overview is Google's own attributable content rather than a display of search results, on pain of a fine up to €250,000 per violation.
Apple's $250M Siri settlement: the price of an over-promised Apple Intelligence, in the public record
Apple agreed to a $250 million settlement in Landsheft v.
Nature Medicine (2026): general-purpose LLMs outperform two deployed specialized clinical AI tools on medical benchmarks
A peer-reviewed Nature Medicine study (Vishwanath et al., 12 June 2026, DOI 10.1038/s41591-026-04431-5) independently evaluated two deployed specialized clinical AI tools, OpenEvidence and UpToDate Expert AI, against three frontier general-purpose LLMs (GPT-5.2, Gemini 3.1 Pro, Claude Opus 4.6) across 500 MedQA questions, 500 HealthBench items and a 100-query real-clinical-query benchmark judged by 12 US clinicians, and found the frontier models outperformed the clinical tools in all three evaluations.
SAVE-O2 AI randomized trial: autonomous oxygen titration kept acutely ill adults in the target range 85% of the time vs 63% under usual care (JAMA Internal Medicine, 2026)
In a peer-reviewed, multicenter randomized trial of 300 acutely ill adults at 4 U.S.
The FCC fined Lingo Telecom $1 million for carrying the AI Biden deepfake robocalls that told New Hampshire voters to stay home
On 21 August 2024 the FCC announced a $1 million settlement with Lingo Telecom, the voice service provider that transmitted spoofed robocalls using a generative-AI clone of President Biden's voice to tell New Hampshire voters not to vote in the January 2024 primary.
An OpenAI reasoning model reanalyzed 376 unsolved rare-disease cases at Boston Children's. Experts confirmed 18 new diagnoses.
In an NEJM AI study, Boston Children's Manton Center, Harvard and OpenAI ran 376 previously unsolved pediatric rare-disease cases through the OpenAI o3 Deep Research model; after expert ACMG/AMP review and CLIA-certified lab confirmation, physicians established 18 new diagnoses, a 4.8% additional yield on cases specialists had already failed to solve.
The FTC fines Cox Media Group and two partners $930,000 over an 'Active Listening' AI ad service that never listened
On 21 May 2026 the FTC announced settlements totalling $930,000 with CMG Media Corporation (d/b/a Cox Media Group), MindSift LLC and 1010 Digital Works LLC over a product branded 'Active Listening', marketed as an AI-powered service that could overhear consumers' conversations through their smart devices and serve targeted ads.
Telefónica closes 2025 with 12 Level 4 autonomous-network use cases, and reports AI cut German network faults by 30%
On 18 February 2026 Telefónica said it had closed 2025 with 12 network use cases running at Level 4 autonomy (self-configuring, self-optimising and self-healing without human intervention), delivered jointly across Spain, Germany and Brazil, inside a program of more than 400 AI use cases in production.
Bartz v. Anthropic: a $1.5 billion settlement, the largest in U.S. copyright history, over pirated books used to train Claude
After Judge William Alsup held in June 2025 that training AI on books was fair use but downloading them from pirate libraries was not, Anthropic settled Bartz v.
When an AI denial tool reaches federal court: in Lokken v. UnitedHealth, the breach-of-contract claims over 'nH Predict' survive Medicare preemption
A putative class action in the U.S.
A German court says you own what your chatbot says: OLG Hamm holds a clinic liable when its AI invented specialist titles
An aesthetic-medicine clinic's website chatbot told visitors its two managing directors were 'specialists in plastic and aesthetic surgery' — a qualification they do not hold, using titles that in two cases do not exist.
Dutch DPA fines Uber €824.99M for automated driver deactivation without human review
On 21 August 2026 the Dutch Data Protection Authority fined Uber €824,990,000, ruling that Uber's software automatically deactivated drivers' accounts between 2018 and 2022 without human assessment — a prohibited fully automated decision under the GDPR.
The UK government banks £45bn a year on AI and digital efficiencies its own auditor says are not yet evidenced
The National Audit Office's July 2026 workforce-planning report records the government's expectation of £45 billion a year in efficiencies from digital transformation and AI, and flags that the published plans do not show how departments derived those savings.
A VA watchdog found errors in nearly all 8,100 automated survivors' death-benefit decisions it reviewed, with $2.7M in improper payments
A VA Office of Inspector General review of some 8,100 automated Dependency and Indemnity Compensation decisions issued over the 12 months through August 2024 found that at least 8,000 contained a legal or procedural deficiency, that at least 2 percent carried legal errors producing an estimated $2,727,764 in improper payments, and that the fault lay in the automation rules themselves - a rule-based process, not artificial intelligence, applying predefined rules without human involvement.
SEC 'AI-washing' enforcement: Rimar Capital and owner Itai Liptz pay $310,000 for an AI trading platform that didn't exist
On 10 October 2024 the SEC announced settled charges against Rimar Capital USA, Rimar Capital LLC, owner Itai Liptz and board member Clifford Boro for false and misleading statements about Rimar's purported use of AI to run automated trading.
A Connecticut federal judge fines a lawyer $3,500 and refers him for discipline over AI-fabricated legal authorities
In Barteca Holdings LLC v.
Pooja Ramesh Singh v. J&K Bank [2026] INSC 668: India's Supreme Court sets aside a tribunal's own AI-hallucinated 'precedents' — 'no decision in the eyes of the law'
In Pooja Ramesh Singh v.
Re A, B, C, D [2026] EWFC 71 (B): a family court names an unregistered barrister who put four AI-hallucinated authorities before it
Sitting at Bournemouth Family Court, Recorder Howard found that Layla Parsons, an unregistered barrister, put four erroneous or fabricated case authorities before the court and accepted she had used a widely known publicly available AI tool to prepare her skeleton argument; the court absolved her of any intention to mislead but held she still did not acknowledge the seriousness, directed that the judgment be published naming her, and recorded her self-report to the Bar Standards Board — with no costs order or fine.
Connecticut's high court, for the first time, sanctions a lawyer for ChatGPT-hallucinated citations
In a per curiam order released July 31, 2026, the Connecticut Supreme Court — confronting a lawyer's use of generative AI 'for the first time' — sanctioned Attorney Ian G.
An AI-drafted petition, a 'wrong draft' excuse, and a referral to the Florida Bar: JMOR Properties v. Artist Alley Townhomes
In JMOR Properties, LLC v.
The Third Circuit's first AI-hallucination sanction: a public reprimand for unverified, AI-generated citations
In its first case addressing attorney misuse of AI, the U.S.
Spain's AEPD fines Mercadona €2,520,000 over its store-entrance facial-recognition system
In procedimiento sancionador PS/00120/2021, terminated by voluntary payment, Spain's data-protection authority fined the grocery chain Mercadona €2,520,000 for its unlawful algorithmic facial-recognition system and prohibited all such processing.
Rechtbank Amsterdam raises a costs award by €2,315.50 after a fabricated DSM/Fox citation the losing party blamed on AI hallucination
In a €2.1m share-purchase dispute, the Rechtbank Amsterdam (ECLI:NL:RBAMS:2026:8258, 12 August 2026) found the defendants' statement of defence quoted the Hoge Raad's landmark DSM/Fox ruling with words found nowhere in that arrest nor in any published case law.
Quebec's appeal court ordered a self-represented litigant to pay $1,000 in costs for filing four ChatGPT-fabricated court decisions
On August 11, 2026 the Quebec Court of Appeal (Sansfacon, Bachand and Harvie, JJ.A.) released reasons in A.
Brazil's ANPD halts Meta's use of personal data to train generative AI (2024)
On 2 July 2024 Brazil's data-protection regulator, the ANPD, issued a Preventive Measure ordering Meta to immediately suspend, in Brazil, its new privacy policy that authorised using users' public data across Facebook, Messenger and Instagram to train generative-AI systems — under a daily fine of R$ 50,000 for non-compliance.
Dutch DPA fines the Tax Administration €2.75M for unlawful algorithmic nationality profiling
On 7 December 2021 the Dutch Data Protection Authority fined the Tax Administration €2.75M for using nationality as an indicator in a self-learning risk-classification algorithm and unlawfully retaining dual-nationality data of childcare-benefit applicants — the algorithmic engine of the toeslagenaffaire scandal.
Hungary's NAIH halts a bank's AI emotion analysis of call recordings, HUF 250,000,000 fine (2022)
On 8 February 2022 Hungary's data-protection authority (NAIH), in decision NAIH-85-3/2022, found that a bank's AI-based analysis of customer-service call recordings, which inferred callers' emotional state to rank them by likely dissatisfaction, seriously infringed the GDPR.
Italy's Garante blocks DeepSeek: limitation on processing of Italian users' data (2025)
On 30 January 2025 Italy's data-protection regulator (the Garante) ordered, urgently and with immediate effect, the limitation on processing of Italian users' personal data by the two Chinese companies behind the DeepSeek chatbot, and opened an investigation, after their response to an information request was deemed inadequate and they claimed EU law did not apply to them.
Italy's Garante fines Foodinho (Glovo) for its rider-management algorithm — €2.6M in 2021, €5M in 2024
Italy's data-protection regulator, the Garante, twice found the rider-management algorithm of Foodinho — Glovo's Italian food-delivery arm — unlawful under the GDPR.
Spain's AEPD orders Worldcoin (Tools for Humanity) to halt biometric iris-scanning data processing (2024)
On 6 March 2024 Spain's data-protection regulator ordered Tools for Humanity Corporation to stop collecting and processing personal data in its Worldcoin iris-scanning project and to block data already collected; on 11 March 2024 Spain's National Court refused Worldcoin's bid to suspend the order.
Adobe updated its Terms of Use after an AI-training backlash to state it won't train generative AI on customer content
In June 2024, an Adobe Creative Cloud Terms of Use update prompted a backlash from creators who read the new language as permission to train AI on their work.
After a December 2023 backlash over a default-on 'third-party AI' setting, Dropbox said customer data is never used to train OpenAI's models and is deleted within 30 days
In December 2023, Dropbox users — including AWS CTO Werner Vogels — discovered a 'third-party AI' toggle turned on by default in account settings and feared their files were being fed to OpenAI as AI-training data.
HubSpot reversed a July 2026 terms-of-service change that would have pooled customer data and allowed AI-model training, after customer backlash
On July 1, 2026, HubSpot's updated terms took effect to support a new 'Contact Discovery' feature and a shared data-enrichment dataset, under which customer data could be used to train HubSpot's AI models unless turned off; independent press reported participation was on by default.
Humane discontinued the Ai Pin and bricked shipped devices; HP bought its AI assets for $116M
On Feb 18, 2025 HP agreed to acquire Humane's Cosmos AI platform, team and 300+ patents/applications for $116 million (forming HP IQ).
A US federal court let a nationwide age-discrimination case against Workday's AI hiring tool proceed as a collective action (May 2025)
On May 16, 2025 U.S.
Mozilla rewrote Firefox's Terms of Use after backlash, saying the change wasn't about using data for AI
In late February 2025 Mozilla gave Firefox its first-ever Terms of Use; a broad data-license clause sparked a backlash and fears the language could feed AI training or ad-targeting.
After a user backlash, Slack updated its AI 'privacy principles' over training ML models on customer data (May 2024)
In May 2024 a Hacker News thread surfaced Slack's published 'privacy principles', which disclosed that Slack analyzes customer messages, content and files to develop AI/ML models on an opt-out-by-default basis, escapable only by email.
X suspended EU/EEA user-data training for Grok after Ireland's DPC took urgent High Court action
On 8 August 2024, X agreed to suspend processing its EU/EEA users' public-post data (collected 7 May – 1 August 2024) to train the AI chatbot Grok, after Ireland's Data Protection Commission brought an urgent High Court application under Section 134 — the first time any EU lead supervisory authority had used such emergency powers.
After backlash, Zoom reversed its terms of service so it will not use customer content to train AI (August 2023)
In March 2023 Zoom changed its online terms of service (effective 27 July 2023); section 10.4 let Zoom use 'Service Generated Data' 'for any purpose ...
WeTransfer removed the machine-learning clause from its Terms of Service after user backlash (July 2025)
After users read a newly added clause 6.3 as permission to train AI on their uploaded files, WeTransfer rolled the change back on 15 July 2025 — removing the machine-learning language and stating it does not use AI to process content shared via the service.
Chicago Sun-Times ran an AI-generated summer reading list where only 5 of 15 books were real
In May 2025 the Chicago Sun-Times published a syndicated summer special section whose 'reading list' recommended books that do not exist; independent newsrooms found only five of the 15 titles were real, the list was AI-generated by a freelancer and licensed via King Features (Hearst), and Chicago Public Media's CEO acknowledged the error and told subscribers they would not be charged.
Google scaled back AI Overviews after it told users to put glue on pizza and eat rocks
In May 2024, days after Google launched AI Overviews, AI-generated answer summaries powered by its Gemini model, to U.S.
Microsoft capped its 'unhinged' new Bing AI chat at 5 turns per session and 50 per day
In February 2023 Microsoft launched a new GPT-powered Bing chatbot.
Microsoft's "Tay" chatbot was pulled offline within ~16 hours after learning to be racist
On 23 March 2016 Microsoft launched Tay, a Twitter chatbot built to emulate a teenage girl and learn from 18-24 year olds.
Zillow shut down its algorithmic home-buying business after a $304M write-down and 25% layoffs
On November 2, 2021 Zillow Group announced it would wind down Zillow Offers, its iBuying operation, which used Zillow's AI home-valuation (the 'Zestimate') to make automated cash offers, taking a ~$304 million Q3 inventory write-down, warning of a further $240-265 million in Q4 losses, and cutting its workforce by ~25% (about 2,000 jobs).
ADGM Court (Heath J.): AED 282,508 indemnity costs against MIO Legal Consultants for AI-fabricated authorities in a defence pleading
In a costs judgment dated 18 December 2025, the Abu Dhabi Global Market Court of First Instance ordered MIO Legal Consultants LLP to pay AED 282,508 on the indemnity basis after its defence relied on unverified AI legal research citing cases that did not exist.
Tribunale di Siracusa n. 338/2026: ~€30,206 in awards after four Cassation citations with fabricated holdings, court infers unverified generative-AI use
An Italian first-instance court (Siracusa, sentenza n.
Arizona federal court: attorney sanctioned for AI-fabricated citations — only 5 to 7 of 19 cases were real, pro hac vice revoked, no fine
In Mavy v.
Miller v. Regions Bank: an Alabama attorney deleted his ChatGPT account to hide AI-fabricated citations — and lost far more than the case
In Miller v.
A federal appeals court fines a lawyer $2,500 for an AI-drafted brief, even though she used purpose-built legal-AI tools
In Fletcher v.
A Massachusetts judge fines a lawyer $2,000 for filing briefs full of AI-invented cases — the state's first
In Smith v.
An Illinois appeals court fined attorney Mason Cole $15,000 for ChatGPT-hallucinated citations
In Scott v.
Lichfield v. Kubler: 10th Circuit finds AI-fabricated citations but declines to sanction
In a published July 27, 2026 opinion affirming the dismissal of Narvin Lichfield's defamation suit against a Netflix documentarian, the U.S.
Ontario Law Society Tribunal orders a lawyer to pay $31,150 in costs after Grok-generated fake citations
In Mazaheri v Law Society of Ontario (2026 ONLSTH 112), the Law Society Tribunal ordered a suspended lawyer to pay the regulator's full costs of $31,150 after he filed motion materials produced by generative AI (which he admitted was 'in particular Grok') that cited non-existent decisions and misused real ones.
OTG New York v. Ottogi America: a New Jersey federal court fines a lawyer $3,000 for a reply brief built on AI-hallucinated citations (2025)
In OTG New York, Inc.
Torres Campos v. Munoz: a published California appellate opinion sanctions a lawyer $5,000 for citing AI-hallucinated cases in a dog-custody fight (2026)
In In re the Domestic Partnership of Torres Campos & Munoz (Cal.
Air Canada held liable for its website chatbot's bad advice: tribunal orders C$812.02 (Moffatt v. Air Canada, 2024 BCCRT 149)
A BC Civil Resolution Tribunal member found Air Canada negligently misrepresented its bereavement-fare policy through the chatbot on its own website, rejected the airline's argument that the chatbot was 'a separate legal entity that is responsible for its own actions', and ordered C$650.88 in damages plus interest and fees, for C$812.02 in total.
A federal court prices an unchecked AI legal workflow: $3,000 sanctions each for the attorneys behind an AI-drafted brief with nearly thirty defective citations
On July 7, 2025, Judge Nina Y.
€30.5 million for an illegal AI face database: the Dutch DPA's final, unappealable GDPR fine on Clearview AI
By decision of 16 May 2024, published 3 September 2024, the Autoriteit Persoonsgegevens (Dutch Data Protection Authority) fined Clearview AI Inc.
iTutorGroup pays $365,000 after the EEOC alleges its application software auto-rejected 200+ older applicants, the settlement the agency's Chair tied to its AI enforcement push
In EEOC v.
FTC final order: the 'AI-powered' web-accessibility widget that promised any-website WCAG compliance pays $1,000,000
On 21 April 2025 the FTC issued final consent order C-4817 against accessiBe Inc.
Federal court enters $25M judgment and permanent ban against Ascend Ecom after FTC alleges its 'AI-powered' passive-income storefronts earned clients virtually nothing
In FTC v.
FTC orders DoNotPay's 'world's first robot lawyer' to pay $193,000 and notify subscribers over unsubstantiated AI claims
The FTC's final Decision and Order (Docket C-4812, issued 14 January 2025, announced 11 February 2025) requires DoNotPay, which advertised its subscription chatbot as 'the world's first robot lawyer,' to pay $193,000 in monetary relief, notify 2021 to 2023 subscribers, and stop claiming its service performs like a real lawyer without evidence.
FTC v. Evolv: the AI scanner marketed as detecting all weapons missed a knife used in a school stabbing, and a federal court entered a permanent injunction
On 5 December 2024 a federal court in Massachusetts entered a stipulated permanent injunction against Evolv Technologies (FTC v.
FTC × IntelliVision: a 'zero bias, millions of faces' AI claim, measured against NIST
IntelliVision sold facial recognition software advertising 'Zero gender or racial bias through model training with millions of faces'.
Federal court bans Rite Aid from AI facial-recognition surveillance for five years after FTC alleges thousands of false-positive matches
A stipulated order in FTC v.
FTC final order: the '98% accurate' AI Content Detector was an unmodified student model, and its own published data showed 53.2% on real-world text
On 28 August 2025 the FTC issued final consent order C-4822 against Workado, LLC (f/k/a Content At Scale AI).
The €15 million ChatGPT fine that didn't survive: Italy's Garante fined OpenAI, the Court of Rome annulled it
By provvedimento n.
A $2.275M class settlement takes the algorithm's score away: Louis v. SafeRent and the five-year rollback of tenant-screening scores for voucher applicants
On 20 November 2024, Judge Angel Kelley of the U.S.
The canonical AI-hallucination sanction: $5,000, jointly and severally, for the lawyers who filed six ChatGPT-fabricated cases in Mata v. Avianca
On June 22, 2023, Judge P.
SEC AI washing enforcement: Delphia and Global Predictions pay $400,000 for AI claims they couldn't back
The SEC's AI washing enforcement actions of 18 March 2024 fined two investment advisers a combined $400,000: Delphia $225,000 and Global Predictions $175,000, after finding their advertised artificial intelligence did not exist as claimed.
SEC AI-washing order: Presto's drive-thru 'voice AI' needed a human on every order
On 14 January 2025 the SEC entered a settled cease-and-desist order against Presto Automation, finding its drive-thru AI claims materially false: the original version of Presto's own voice AI 'required human agent intervention, including entering the order, in all instances' (off-site agents in the Philippines and India) while filings claimed it 'eliminat[es] human order taking' and reported 95% to 99% automation rates that counted only restaurant-staff involvement.
A federal court fined a public defender $1,500 for a fake citation, then declined to say AI wrote it
United States v.
John Deere See & Spray: independent university trials cut post-emergence herbicide 43% to 62%
Computer-vision targeted spraying, measured not by John Deere but by independent university weed scientists.
Leiske v. Kidd (Del. Ch.): one of Delaware's oldest firms — Richards, Layton & Finger — ordered to show cause over GenAI-hallucinated citations. No sanction yet.
On 2026-07-01 Vice Chancellor Lori W.
Romero v. Corona Investments: the Seventh Circuit found 'an astonishing number of hallucinated citations' — and declined to sanction anyway
In In re Romero (Romero v.
Akerlund v. Atlas Air: a federal appeals court refers a former Florida lawmaker to its lawyer-conduct committee over AI-hallucinated citations, filed twice
In Akerlund v.
The UK's DWP fraud model: a machine-learning system that saved an estimated 4.4 million pounds on Universal Credit advances
Since May 2022 the UK Department for Work and Pensions has run a machine-learning model that flags potentially fraudulent Universal Credit advance claims for review.
Meta pulls Instagram's Muse Image @-mention feature three days after launch
Meta launched a Muse Image feature on July 7, 2026 that let anyone generate AI images by @-mentioning public Instagram accounts - on by default for adults.
Coinbase: AI spend cut nearly in half, and the biggest lever was caching, not cheaper models
In late June 2026, Coinbase CEO Brian Armstrong said the company cut internal AI spending by nearly half while token usage sat near a company high.
Notion and Decagon: an AI support agent, 34% faster resolution and 2x deflection
Notion deployed Decagon's AI customer-experience agent to automate repetitive inquiries and route the rest to the right expert.
Delivery Hero and n8n: account-lockout recovery automated in a single workflow
Delivery Hero's global IT service delivery team automated employee account-lockout recovery with one n8n Enterprise workflow: the employee's manager approves, and API calls to Okta, Jira, and Google restore access.
JPMorgan's COIN: the famous 360,000-hours figure, and why it was never independently measured
JPMorgan built COIN (Contract Intelligence), an in-house machine-learning system that interprets commercial-loan agreements in seconds, work that Bloomberg reported in 2017 had consumed 360,000 hours a year by lawyers and loan officers.
Klarna's AI customer-service assistant: the 2024 numbers and the 2025 walk-back
Klarna's OpenAI-powered assistant handled two-thirds of customer-service chats in its first month (Feb 2024) and was said to do the equivalent work of 700 full-time agents, a modeled equivalence, not 700 layoffs.
Mango's first fully AI-generated campaign: Sunset Dream, 95 markets, and no public performance number
In July 2024 Mango ran what it calls its first campaign generated entirely with generative AI, for the Sunset Dream collection of its Mango Teen line, live in 95 markets.
Moderna's company-wide ChatGPT Enterprise rollout: 750 custom GPTs, from a vendor case study
Moderna deployed OpenAI's ChatGPT Enterprise company-wide in April 2024 and reported 750 custom GPTs within two months, 120 conversations per user per week, and 100% adoption in its legal team.
Octopus Energy's 'Magic Ink': AI that drafts support emails, with no job cuts
Octopus Energy's Kraken 'Magic Ink' drafted customer-service email replies with a human reviewing and sending each one.
Omega Healthcare and UiPath: document automation across billing, coding, and payer correspondence
Omega Healthcare, a healthcare revenue cycle management firm, has run UiPath automation for about five years across billing, medical coding, and insurance correspondence.
Otto's autonomous stock ordering: a deep-learning system that buys inventory on its own
German retailer Otto uses a Blue Yonder deep-learning system, built on an algorithm that originated at CERN, to forecast demand and automatically order stock.
PZU and Tractable: from a detailed review of 20% of body-shop claims to nearly all, in real time
PZU, the largest insurer in Central and Eastern Europe, deployed Tractable's computer-vision AI to review motor-damage claims from body shops.
TSXG Galicia: Spain's High Court of Justice of Galicia fines a lawyer €1,800 for an appeal built on 24 AI-fabricated citations, and sizes the fine to a legal-AI subscription
The Sala de lo Social of the Tribunal Superior de Xustiza de Galicia fined a lawyer €1,800 under article 75.4 LJS after his social-jurisdiction appeal was found to contain 24 fabricated jurisprudential citations (of 32, 75%), attributed to unverified generative-AI 'hallucinations', and referred him to the A Coruña Bar Association.
Zalando's 2018 'algorithms replace 250 marketing jobs': the headline, and the hiring plan it left out
In March 2018 Zalando said it would replace 200 to 250 marketing roles with algorithms and AI, a line cited for years as a clean AI-for-layoffs example.
Landberg v. City of New York: an AI-fabricated appellate brief drew $10,500 in sanctions
A New York appellate court (Second Department) sanctioned attorney Michael Sanders $8,000 and his law firm $2,500, $10,500 total, to the Lawyers' Fund for Client Protection, after an appellant's brief prepared with generative AI cited three nonexistent cases, fabricated Court of Appeals quotations contrary to the actual law, and misrepresented real cases (2026 NY Slip Op 03935, decided June 23, 2026).
Italy's Supreme Court fines an appellant €5,000 for AI-hallucinated case citations
In judgment no.
Barber v. Morawa: Michigan's first published appellate opinion sanctioning a lawyer for AI-fabricated citations
In Barber v.
Withers v. City of Aberdeen: a federal court sanctions attorneys on both sides $8,000 for AI-hallucinated citations
In a June 8, 2026 Sanctions Order, U.S.
Alabama Supreme Court dismisses appeal, fines lawyer $17,200 for AI-'hallucinated' citations
In Ibach v.
Rodney v Gee'z Micro Bar: a County Court refers two solicitors to the SRA over AI-suspected fake citations
In an appeal at the County Court at Dudley, His Honour Judge Grimshaw found that three documents filed by the firm AML Legal contained incorrect and misleading citations of legal authorities with the suspicion they had been created using AI, held that admonishment alone was insufficient, and referred both the firm's director and the responsible consultant solicitor to the Solicitors Regulation Authority — declining contempt proceedings and imposing no monetary sanction.
United States v. Farris: a federal appeals court denies a lawyer his pay and removes him from the case over AI-fabricated quotations from Westlaw's CoCounsel
In United States v.
Italy's Supreme Court fines an appellant EUR 3,000 for AI-hallucinated citations — Cass. pen. Sez. 7 Ord. 11431/2026
In Ordinanza n.
Prososki v. Regan: Nebraska Supreme Court strikes an appellate brief and refers counsel for discipline over fictitious, allegedly AI-generated citations
In a per curiam opinion filed March 20, 2026, the Nebraska Supreme Court struck the appellant's brief, dismissed the appeal, and referred appellant's counsel, W.
Whiting v. City of Athens: a federal appeals court fines two lawyers $15,000 each for briefs with 'over two dozen fake citations'
In consolidated appeals over a 2022 Tennessee fireworks show, the U.S.
In re Nwaubani: the Fourth Circuit publicly admonishes a lawyer over nonexistent citations 'potentially derived from generative AI'
The U.S.
Ars Technica retracts an article after an AI tool fabricated quotations
On Feb 13, 2026, Ars Technica retracted in full a story that carried AI-fabricated quotations falsely attributed to matplotlib maintainer Scott Shambaugh; the editor-in-chief called it 'a serious failure of our standards' and apologized.
Amarsingh v. Frontier Airlines: 10th Circuit sanctions attorney $1,000 for ChatGPT-fabricated citations
In Amarsingh v.
A New York court annuls a university's AI-detection finding — and orders the record expunged
In Matter of Newby v Adelphi University, the New York State Supreme Court annulled an academic-integrity finding built on a Turnitin '100%' AI-detection score, calling it 'without valid basis and devoid of reason' and ordering the record expunged.
Kjoller v. Superior Court: California's high court steps in on a prosecutor's AI-fabricated brief
On 2026-01-14 the California Supreme Court granted-and-transferred Kjoller v.
New York's first appellate AI-sanction: a lawyer who conceded using AI, 23 fabricated cases, and $10,000 in penalties (Deutsche Bank Natl. Trust Co. v LeTennier)
In Deutsche Bank Natl.
Dastou v. Holmes: a Massachusetts judge sanctions a lawyer for ChatGPT-fabricated jury instructions, no fine, but an MCLE course and no billing the client
In Dastou v.
Couvrette v. Wisnovsky (D. Or.): a $15,500 sanction for AI-fabricated citations, at $500 per fake case and $1,000 per fabricated quotation
A federal magistrate judge in Oregon sanctioned plaintiffs' counsel Stephen Brigandi $15,500 for filing three summary-judgment briefs that cited fifteen non-existent cases and eight fabricated quotations, adopting a $500-per-case / $1,000-per-quotation formula, then struck the briefs, dismissed the case with prejudice, and referred counsel to the Oregon State Bar.
A federal judge struck a lawyer's brief over 17 AI-fabricated quotations and folded the conduct into a grievance referral
In the consolidated Fat Joe litigation (Dixon v.
Oregon Court of Appeals fines a lawyer $10,000 for AI-fabricated authority: the state's largest such penalty and the Ringo formula's biggest bill yet
In Doiban v.
Fivehouse v. U.S. Dept. of Defense (E.D.N.C.): a federal prosecutor is publicly reprimanded for a brief with AI-fabricated quotations
A federal magistrate judge in North Carolina publicly reprimanded former Assistant U.S.
A federal court fined a lawyer $6,000 for a brief built on an AI-hallucinated case
In Gardner v.
Gauthier v. Goodyear: a Texas lawyer is sanctioned $2,000 for filing AI-hallucinated case citations
In a wrongful-termination suit, plaintiff's counsel used a generative-AI tool to draft a brief and did not verify it.
In re Marla C. Martin (Bankr. N.D. Ill.): $5,500 Rule 9011 Sanction for a ChatGPT-Fabricated Brief
A federal bankruptcy judge in Chicago found that debtor's counsel Thomas E.
A federal judge fined a lawyer $5,000 for filing AI-fabricated case citations against T.D. Jakes
In Jakes v.
Johnson v. Dunn: three lawyers are disqualified and referred to the bar for filing five ChatGPT-hallucinated citations
Three Butler Snow attorneys defending an Alabama prison official filed two motions with case citations invented by ChatGPT, five made-up citations in all.
An Ohio federal judge fines two lawyers $7,500 for AI-fabricated citations, 'the most egregious' Rule 11 violation in his 46 years on the bench
In Kettering Adventist Healthcare v.
Kohls v. Ellison: a court throws out a Stanford misinformation expert's declaration after GPT-4o invented its citations
To defend Minnesota's political-deepfake law, the state's Attorney General filed an expert declaration from a Stanford AI-misinformation scholar.
Kruse v. Karlen: Missouri appeals court fines a pro se litigant $10,000 after 22 of 24 cited cases turn out to be AI fabrications (Mo. App. E.D., Feb. 2024)
A non-lawyer's AI-drafted appellate brief cited 24 cases; only two were real.
Nine of 27 citations were wrong and two cases did not exist: a special master ordered Ellis George and K&L Gates to pay $31,100
In Lacey v.
A Kansas federal judge fined five attorneys $12,000 for AI-fabricated citations, and named the failure, not the AI
In Lexos Media IP, LLC v.
The Ninth Circuit fines two lawyers $2,500 each and suspends them for six months over AI-hallucinated briefs, but the sanction is for candor, not for using AI
In Lnu v.
A federal court fined a lawyer $6,000 for three briefs full of AI-fabricated citations
In Mid Central Operating Engineers v.
Oregon Court of Appeals fines a lawyer $2,000 for AI-fabricated authority, and sets a per-item price: $500 a fake citation, $1,000 a fake quote
In Ringo v.
Wadsworth v. Walmart: AI hallucinated eight of nine cases in a Morgan and Morgan brief, and a federal judge fined the lawyers $5,000 and revoked one pro hac vice
In Wadsworth v.
Czech Constitutional Court fines a lawyer 25,000 CZK for a complaint built on AI-hallucinated case law
The Constitutional Court of the Czech Republic imposed a 25,000 CZK procedural fine on Prague lawyer Mgr.
Getty Images v Stability AI [2025] EWHC 2863 (Ch): UK High Court rejects the secondary copyright claim and finds only 'historic' and 'extremely limited' trade-mark infringement
On 4 November 2025 the England and Wales High Court handed down Getty Images (US) Inc & Ors v Stability AI Limited [2025] EWHC 2863 (Ch): Getty abandoned its primary copyright claims during trial, the court rejected the remaining secondary copyright claim because Stable Diffusion's model weights do not store the works, and the only infringement found was 'historic' and 'extremely limited' trade-mark use of the Getty watermark in early model outputs.
Character.AI removes open-ended chat for under-18 users, retreating from the AI-companion model
On 29 October 2025 Character.AI announced it will remove the ability for users under 18 to engage in open-ended chat with AI on its platform, the core of its companion product, taking effect no later than 25 November 2025, with an interim two-hours-per-day cap ramping to zero.
United States v. McGee: a federal judge fines a court-appointed defense lawyer $5,000 for AI-hallucinated citations produced by a Microsoft Word plug-in called Ghostwriter Legal (2025)
In United States v.
Noland v. Land of the Free: California's first published opinion on AI-fabricated citations ($10,000 sanction)
In California's first published appellate opinion on AI-fabricated citations, the Court of Appeal (2d Dist., Div.
AI-fabricated quotes and nonexistent citations in a Supreme Court of Victoria murder case
Defence King's Counsel Rishi Nathwani apologised to Justice James Elliott after filing submissions containing AI-generated fabrications — fake quotes from a legislature speech and nonexistent case citations — in a Melbourne murder proceeding.
Ayinde v Haringey / Al-Haroun v Qatar National Bank: the UK High Court's landmark ruling on AI-fabricated case-law
In two joined cases, the England & Wales Divisional Court (Dame Victoria Sharp P and Johnson J) sanctioned lawyers who put fabricated, AI-generated (or suspected-AI) case authorities before the courts — a pupil barrister and a law centre were ordered to pay £2,000 each in wasted costs and referred to the Bar Standards Board and the Solicitors Regulation Authority, and the court warned that misuse of AI could amount to contempt.
Garner v. Kadince: Utah Court of Appeals Sanctions Counsel for ChatGPT-Hallucinated Citations (2025 UT App 80)
In Utah's first published appellate opinion on AI-hallucinated citations, the Utah Court of Appeals sanctioned Petitioner's counsel under rule 40 for filing a petition drafted with ChatGPT that cited a non-existent case, 'Royer v.
Murray (Wamba Wemba) v Victoria: a Federal Court of Australia indemnity-costs order after AI-fabricated citations
In a native title case, Justice Bernard Murphy of the Federal Court of Australia ordered the applicant's solicitors, Massar Briggs Law, to personally pay the respondents' costs on an indemnity basis after two court documents contained false and incorrect citations that the court found seemed likely to have arisen through Generative AI; a junior solicitor had used the Google Scholar search tool and failed to verify the results.
Thomson Reuters v. Ross Intelligence: Delaware court rejects fair use for AI training data, finds 2,243 Westlaw headnotes infringed
On February 11, 2025, U.S.
Meta paused training its AI (LLMs) on EU/EEA Facebook and Instagram public data after the Irish DPC asked it to delay (June 2024)
On 14 June 2024 Meta paused its plan to train its generative-AI large language models on the public Facebook and Instagram content of EU/EEA adults, after its lead EU regulator, the Irish Data Protection Commission, asked it to delay; Meta said it was 'disappointed by the request,' and the DPC said it 'welcomes the decision by Meta to pause its plans.'
AI-ECG mortality alert: a 15,965-patient randomized trial in Taiwan cut 90-day all-cause mortality from 4.3% to 3.6% (HR 0.83)
In a multisite pragmatic randomized controlled trial at Tri-Service General Hospital (National Defense Medical Center, Taiwan), a deep-learning ECG classifier flagged high-mortality-risk inpatients and pushed an alert to their physicians.
ICO orders Serco Leisure to halt unlawful facial-recognition and fingerprint attendance monitoring of 2,000+ staff (2024)
On 23 February 2024 the UK Information Commissioner's Office ordered Serco Leisure, Serco Jersey and seven associated community leisure trusts to stop using facial-recognition technology and fingerprint scanning to monitor the attendance of more than 2,000 employees at 38 leisure facilities, finding the biometric processing unlawful under the UK GDPR and ordering the data destroyed within three months.
No files match. Loosen a filter or clear the search.
Have proof worth filing?
If you delivered a real automation or AI-ops outcome, bring it to the dojo and earn a place in the record.