CASE STUDIES
Global AI Case Repository
A curated repository of landmark judicial decisions and cutting-edge governance practices worldwide, analyzing judicial reasoning and compliance standards across technical scenarios.
Global Case Map
Judicial Precedents across Global Jurisdictions
Covering global AI judicial decisions and governance practices, searchable by jurisdiction, region, and core legal issues.
Select a region on the map to filter corresponding judicial cases.
China base map: Ministry of Natural Resources standard map service · View source map
Showing all indexed cases
Beijing AI Text-to-Image Case
The court found the disputed image reflected the user’s intellectual input and individual expression, considering prompts, parameter adjustments and the generation process. Protection depended on the specific human contribution in this case; the decision does not make every AI-generated image copyrightable.
Effective first-instance judgmentShanghai AI Face-Swapping Case
A mini-program provided paid generation services by replacing facial features in original videos. The court held that localized face-swapping did not constitute an original adaptation and that the platform's commercial use infringed the right of dissemination through information networks.
Effective first-instance judgmentGuangzhou Ultraman AI Image Case
The defendant connected to a third-party AI image service through an API. Users could generate images substantially similar to the protected Ultraman works. The court classified the defendant as a generative AI service provider and found infringement of the reproduction and adaptation rights.
First-instance judgmentUS Thaler Copyright Case
The applicant named an AI system as the sole author. The appellate court upheld refusal under the human-authorship requirement of the Copyright Act of 1976, and the Supreme Court later denied certiorari. The case did not decide copyrightability of AI-assisted work with substantial human authorship.
Registration refusal affirmed; Supreme Court denied certiorari on 2 March 2026US Attorney Fictitious Citation Case
Attorneys submitted court filings containing fictitious judicial decisions and citations fabricated by generative artificial intelligence, and continued to provide inaccurate materials after being challenged. The court imposed sanctions accordingly.
Sanctions order against lawyers and their firmUK DABUS Patent Case
The applicant named DABUS as the sole inventor of two inventions and claimed entitlement through ownership of the system. Under the Patents Act 1977, the Supreme Court unanimously held that an inventor must be a natural person and system ownership alone did not confer entitlement.
Final judgment of the UK Supreme CourtUK Getty Images Case
Getty challenged training, model distribution, outputs and trademark use involving Stable Diffusion. It abandoned direct copyright and database-right claims at trial without a merits determination. The High Court rejected secondary copyright infringement and found limited trademark infringement in watermark outputs of specified older versions.
High Court judgment; Getty permitted to appeal secondary copyright rulingAir Canada Chatbot Case
Air Canada’s website chatbot incorrectly stated that bereavement fares could be requested retroactively within 90 days of ticket issuance. Applying negligent misrepresentation principles, the tribunal found liability based on the airline’s duty of care, the consumer’s reasonable reliance and resulting loss.
Civil Resolution Tribunal small-claims decisionEU SCHUFA Scoring Case
A credit agency automatically generated credit scores for use by third parties such as banks. The Court of Justice of the European Union held that when a score plays a decisive role in a third party's decision, the scoring activity constitutes automated individual decision-making governed by Article 22 of the GDPR.
Preliminary-ruling judgment interpreting EU lawAustralia: Suspected AI-Assisted Character Reference
At sentencing, the judge suspected the defendant’s brother’s character reference had used a large language model or machine translation and gave it very little weight. Its AI origin was not established.
Sentencing JudgmentBeijing AI Voice Case
Recordings licensed for conventional use were repurposed into an AI voice product. The court protected personality interests in a recognizable voice, assessing whether listeners could associate the synthesized output with the particular person.
Effective JudgmentU.S. ROSS Training Data Case
ROSS used materials derived from Westlaw headnotes to train a competing legal research tool. On September 29, 2026, the Third Circuit affirmed the district court’s rulings on originality and fair use. The decision concerns a non-generative AI system and an interlocutory appeal; it does not resolve every remaining issue in the litigation.
Third Circuit affirms partial summary judgmentU.S. Anthropic Book Training Case
On the specific facts, the court distinguished model training, one-for-one digitisation of purchased books with destruction of the originals, and a library acquired from pirated sources. It later finally approved a USD 1.5 billion class settlement of defined past input and copying disputes; approval did not license future conduct.
Class settlement finally approved on 20 July 2026; implementation underwayU.S. Meta Book Training Case
On evidence submitted by 13 authors, the court entered summary judgment for Meta on copying books to train Llama. It did not decide every distribution, contributory-infringement or proposed class issue.
Summary judgment on training-copying claims; distribution and contributory claims remainU.S. Workday Hiring Algorithm Case
Job applicants allege that Workday’s screening tools caused discriminatory outcomes. Earlier rulings allowed claims under an employer-agent theory and the ADEA. On June 22, 2026, the court allowed most FEHA claims and Hughes’s amended ADA claim to proceed, while dismissing certain claims added without leave. These pleading-stage rulings do not establish that the tools actually caused unlawful discrimination.
Motion to dismiss granted in part and denied in part; remaining claims in discoveryUS COMPAS Sentencing Case
Criminal sentencing materials incorporated a commercial COMPAS recidivism risk score. The court permitted limited consideration of the score, required clear warnings regarding the tool's limitations, and prohibited using the score to determine sentence severity or whether incarceration is warranted.
State Supreme Court judgment affirming denial of resentencingUS Lovo Voice Cloning Case
Voice actors allege their recordings were used beyond authorization for voice cloning. The 10 July 2025 ruling preserved contract, New York personality-use, consumer-protection and specified claims about use of Sage’s original recordings. Training-copying claims were dismissed with leave to amend for insufficient factual pleading. The defendant later filed a bankruptcy notice; there was no final merits finding of liability.
Partial dismissal ruling; defendant filed a bankruptcy notice in May 2026; later procedure unverifiedEU Automated Decision-Making Explanation Case
Interpreting GDPR Article 15(1)(h), the CJEU required a concise, transparent and intelligible account of the procedures and principles actually used, enabling the individual to understand and challenge the result. Showing how changes in personal data would change the result is one possible explanatory method.
Preliminary-ruling judgment interpreting EU lawNetherlands SyRI Risk Profiling Case
The Dutch government integrated multi-agency public data to generate fraud risk profiles for social welfare recipients. The court held that the regulatory regime lacked sufficient transparency and auditability, failing to strike a fair balance in its interference with the right to private life.
First-instance civil judgment; government did not appealNetherlands Ola Driver Data Access Case
UK drivers sought personal data and explanations concerning Ola’s scoring, fraud detection, income profiles, allocation and deductions. The first-instance court partly allowed access in 2021. In 2023 the Amsterdam Court of Appeal found that the fraud-probability scores, income profiles and allocation at issue had significant effects and ordered further explanations.
First-instance order; partly reversed on appeal on 4 April 2023UK Police Live Facial Recognition Case
South Wales Police deployed a live facial recognition system in public spaces. The Court of Appeal held that the criteria for watchlists and deployment locations lacked sufficient legal constraints, and identified deficiencies in the data protection impact assessment and compliance with the public sector equality duty.
Appellate JudgmentEwert v. Canada Risk Assessment Case
The Supreme Court of Canada held that the correctional authority breached its statutory duty by failing to reasonably validate psychometric and actuarial risk tools for Indigenous people. It did not establish empirical bias or uphold all Charter claims. The case is used as a comparative example of algorithm governance.
Supreme Court JudgmentJapan DABUS Patent Case
The application named DABUS as inventor. Japan’s IP High Court affirmed rejection under the natural-person inventorship requirement. Counsel reported the later Supreme Court refusal; the Supreme Court order was not independently obtained.
IP High Court affirms rejection; counsel reports the Supreme Court declined the petition on 4 March 2026Singapore Algorithmic Trading Contract Case
After deterministic trading programs executed trades at an abnormal rate, the platform reversed them. The court found breach of contract and considered programmers’ or operators’ relevant knowledge from programming through contract formation, including later awareness of an error followed by allowing the program to exploit it.
Final JudgmentColombia Judicial Use of ChatGPT Case
A lower-court judge included ChatGPT responses in the reasoning of a judgment. The Constitutional Court found that the judge had independently decided the matter prior to querying the model, upholding the validity of the proceedings while establishing standards for judicial AI use including transparency, verification, privacy, and human control.
Constitutional Court JudgmentSouth Africa: Fabricated Citations and Verification Duties
After fabricated authorities were submitted, the court refused leave to appeal, required the firm to bear the costs of two additional hearings and referred the matter to the Legal Practice Council. The court discussed AI research risks but did not establish the source of the erroneous citations.
High Court JudgmentIndia ANI v. OpenAI Case
On 24 July 2026, the Delhi High Court dismissed ANI’s interim-injunction application on the record before it. Its prima facie conclusions on fair dealing and alleged output infringement do not determine the merits at trial.
Interim injunction refused; merits unresolved and latest appeal status unverifiedFrance Parcoursup Algorithm Case
A student organization sought disclosure of a university’s local Parcoursup algorithm and source code. The Conseil d’État recognized rejected applicants’ individual explanation rights and post-admissions publication of criteria and algorithm use, without requiring disclosure of the algorithm or source code to third parties.
Conseil d’État judgment: no need to adjudicate part; remaining claims dismissedSpain Glovo Couriers Case
Glovo organized delivery services through its platform, rating systems, automated order dispatch, and geolocation tracking. The Supreme Court of Spain held that an employment relationship existed between the courier and the platform, citing algorithmic management and the platform's control over essential assets.
Supreme Court JudgmentBrazil: Metro Facial Detection and Emotion Analysis
Metro advertising equipment detected passenger age, gender and emotions. The appellate court increased collective moral damages to BRL 500,000 and upheld restrictions on processing without consent. The case concerned facial detection and emotion analysis, without establishing identity matching.
Appellate JudgmentBeijing AI Voice Endorsement Case
A book retailer used a public figure's image and a highly similar AI-synthesized voice in promotional videos. The court found infringements of portrait and voice rights and held that a merchant commissioning influencer promotion must review the promotional content.
Effective first-instance judgment; republished in the SPC’s September 2026 representative casesBeijing AI Face-Swap Personal Information Case
A face-swap application turned videos featuring two online models into paid templates. The court distinguished portrait identifiability from the processing of facial and other personal information, finding unauthorized personal information processing.
Effective judgment; released as a typical case in September 2025Beijing AI Content Detection Case
The platform hid a user’s contemporaneously written answer and imposed a one-day posting restriction based on algorithmic detection. Because the platform controlled the detector and its reasoning, it owed a reasonable, proportionate explanation on these facts; the unsupported action breached the contract.
First-instance judgment effective after withdrawal of the appeal; published as a representative case in September 2025Beijing AI Portrait Parody Case
The court distinguished the group publication, which infringed portrait and reputation rights, from later private messages that violated the claimant’s dignity and general personality rights.
Effective judgment; released as a typical case in September 2025Beijing Virtual Human Copyright Case
The protected work was the human-designed appearance of a virtual character. The courts assessed the uploader and platform separately; the digital character was not itself a copyright holder.
Judgment upheld on appeal and effective; published as a representative case in September 2025Beijing AI Companion Personality Case
A mobile application allowed users to create AI companions using a public figure's name, portrait, and persona, while the operator organized, recommended, and expanded those characters through algorithms. The court found infringement of the public figure's personality rights.
First-instance judgment effective after withdrawal of the appeal; published as a representative case in September 2025UK AI False Citation Review
Under the Hamid jurisdiction, the court jointly examined fabricated authorities and verification duties in two proceedings, without merging their underlying merits. Al-Haroun admitted using AI tools and other sources; the Ayinde lawyer denied generative AI use, and the court expressed serious doubts without establishing a unique cause.
High Court judgment reviewing lawyers’ professional responsibilityU.S. Character.AI Product Liability Case
A mother brought product-liability and negligence claims concerning an AI companion chatbot after her teenage son’s death. A May 2025 ruling allowed several claims to proceed beyond dismissal. After the parties reported a settlement, the court dismissed and closed the case in January 2026 without a final determination of liability on the merits.
Dismissed and closed following notice of settlement on January 7, 2026Shenzhen AI-Assisted Adjudication System
Shenzhen courts have developed an AI-assisted adjudication system and a vertical model for judicial work. Public reports describe its use in assisting more than 600,000 cases and illustrate governance needs for judicial data, access permissions, accuracy, and human responsibility.
Public judicial application caseShanghai RAG Search Links Case
An AI search engine returned links to unauthorised television copies. On the specific facts, the courts found neither direct nor contributory infringement by the operator.
Dismissal upheld on appeal; published in an SPC representative-case release in 2026Zhangjiagang AI Butterfly Chair Case
A designer lacked original generation records for the disputed AI images. The court found insufficient evidence of original human contributions to their expression and dismissed the claims.
First-instance judgment effective after withdrawal of the appeal; official representative-case reportHangzhou AI Face-Swapping Public-Interest Case
Unauthorised face-swapping for sexually explicit content prompted public-interest litigation. The court ordered deletion, an apology and public-interest damages.
Effective civil public-interest judgment; compliance reported publiclyBeijing AI Speech Dataset Case
An openly available speech dataset lacked compilation copyright but supported protected competitive interests. The appellate court applied unfair-competition law and treated registration as preliminary evidence.
Concluded with an effective judgment in 2024; damages upheld on appeal on a different legal basis; annual representative-case release on 30 April 2025, not the first reportHangzhou Ultraman LoRA Platform Case
Users trained and shared Ultraman LoRA models and images. Identifiable risks, available intervention and direct revenue supported a finding of contributory infringement.
Appeal dismissed and judgment upheld; Zhejiang High People’s Court representative-case reportHangzhou AI Hallucination Liability Case
A general-purpose chatbot gave incorrect admissions information and generated a compensation promise. The court rejected the claim after assessing attribution, care, harm and causation.
First-instance judgment effective in 2025; court-origin case reportingNo cases match current filters
Consider adjusting your search terms, jurisdictions, or issue filters.
Judicial summaries and practical analyses are prepared for professional research. Specific engagements should verify the latest official texts, appellate developments, and applicable statutory laws.