Generated Source Mirror¶
This page is generated from the source Markdown file so it can be viewed inside the MkDocs site.
- Source path:
knowledge/legal/TOPIC-AI-LITIGATION.md - Source ID:
TOPIC-AI-LITIGATION - Edit this page | Suggest a change
📝 Page Status: Draft — sourced but not yet reviewed
- Status:
draft - Sources: 14
- Relationships: 16
- Research debt items: 5
AI Litigation Landscape¶
Summary¶
Generative AI is now a regular subject of civil litigation. The cases documented here fall into four clusters: (1) copyright in training data and model outputs, including news, books, and music, (2) copyright and trademark claims over AI-generated viral characters, (3) product-liability and speech defenses when chatbot products are alleged to have harmed users, and (4) one comparative German first-instance lyric judgment. This page is an inventory, not a prediction of outcomes.
Verified Facts¶
- Thaler v. Perlmutter (D.C. Cir. Mar. 18, 2025) held that the Copyright Act requires human authorship and affirmed refusal to register a work attributed solely to a generative-AI system.1
- The Copyright Office’s January 2025 copyrightability report concluded that purely AI-generated material is not copyrightable and that prompts alone do not provide sufficient human control on current generally available technology.2
- Thomson Reuters v. Ross Intelligence (D. Del. Feb. 11, 2025) rejected fair use for a non-generative legal-research AI trained on Westlaw headnotes to compete with Westlaw.3
- Bartz v. Anthropic PBC (N.D. Cal. June 23, 2025) held that training Claude on books was transformative fair use but that a pirated permanent library was not; a $1.5 billion class settlement of piracy claims was finally approved July 20, 2026.45
- Kadrey v. Meta Platforms, Inc. (N.D. Cal. June 25, 2025) granted Meta fair-use summary judgment on Llama training as to thirteen authors on that record, while leaving torrenting-distribution claims for later proceedings.6
- The New York Times Company v. Microsoft Corp. (S.D.N.Y.) survived most of an April 4, 2025 motion to dismiss; an August 6, 2026 order dismissed material-contribution contributory claims premised on alleged end-user infringement.78
- Spyder Games LLC v. Mementum Lab (N.D. Cal., filed Nov. 26, 2025) seeks a declaration that AI-generated “brainrot” characters used in Steal a Brainrot are not copyrightable. See Brain Rot and AI-Generated Meme Characters.9
- Garcia v. Character Technologies, Inc. (M.D. Fla. May 20, 2025) allowed most wrongful-death and product-liability claims against a chatbot company and Google to proceed and declined, at the pleading stage, to treat LLM output as protected speech.10
- Music copyright dockets are inventoried in Music Copyright and Generative AI: UMG v. Suno and UMG v. Udio (sound recordings; partial dismissals; no U.S. fair-use holding), Concord v. Anthropic (lyrics; secondary and DMCA claims survive dismissal), and GEMA v. OpenAI (non-final German first-instance lyric judgment).11121314
Case map¶
| Cluster | Leading OIR page | What is actually decided (as of sources cited) |
|---|---|---|
| Output authorship | Thaler | Machine listed as sole author cannot register. |
| Training vs piracy | Bartz | Training fair use; pirated library not; piracy claims settled. |
| Training market harm | Kadrey | Fair use on this record; dilution theory undeveloped. |
| Competing research AI | Ross | Not fair use; tool was not generative. |
| News outputs | Times v. OpenAI | Claims largely survive dismissal; some secondary theories dismissed. |
| Meme-character ownership | Spyder | Pending declaration action. |
| Music masters | Suno / Udio | Pending; partial dismissals; no U.S. fair-use holding. |
| Music lyrics (U.S.) | Concord | Secondary and CMI claims survive 12(b)(6). |
| Music lyrics (DE) | GEMA | First-instance liability; not final. |
| Chatbot harm | Garcia | Design-defect claims proceed; speech defense not established at 12(b)(6). |
Legal Analysis¶
These dockets are not one “AI law.” Copyright training cases apply § 107 to copying that happens before a user sees an answer. Authorship cases ask who, if anyone, owns the answer or the image. Product-liability cases ask whether the software is a product whose design can be defective, which collides with First Amendment arguments that outputs are speech.10
For open-internet and software-freedom work, two tensions matter. First, a broad training-is-always-infringement rule would raise the cost of building models, including open-weight models; Bartz and Kadrey show district judges already disagree about how to weigh that against market harm.46 Second, treating chatbot text as unprotected non-speech at the pleading stage, as in Garcia, could expand design-defect litigation against conversational software without a merits record on causation.10
Criminal and sanctions cases involving software authors remain on Documented Proceedings Involving Software Authors and Operators; they are a different inventory.
Relationships¶
TOPIC-AI-LITIGATIONrelated_toTOPIC-AI-COPYRIGHT.TOPIC-AI-LITIGATIONrelated_toTOPIC-BRAIN-ROT.TOPIC-AI-LITIGATIONrelated_toCASE-THALER-V-PERLMUTTER.TOPIC-AI-LITIGATIONrelated_toCASE-BARTZ-V-ANTHROPIC.TOPIC-AI-LITIGATIONrelated_toCASE-KADREY-V-META.TOPIC-AI-LITIGATIONrelated_toCASE-THOMSON-REUTERS-V-ROSS.TOPIC-AI-LITIGATIONrelated_toCASE-NYT-V-OPENAI.TOPIC-AI-LITIGATIONrelated_toCASE-SPYDER-V-MEMENTUM.TOPIC-AI-LITIGATIONrelated_toCASE-GARCIA-V-CHARACTER-TECHNOLOGIES.TOPIC-AI-LITIGATIONrelated_toTOPIC-AI-MUSIC-COPYRIGHT.TOPIC-AI-LITIGATIONrelated_toCASE-UMG-V-SUNO.TOPIC-AI-LITIGATIONrelated_toCASE-UMG-V-UDIO.TOPIC-AI-LITIGATIONrelated_toCASE-CONCORD-V-ANTHROPIC.TOPIC-AI-LITIGATIONrelated_toCASE-GEMA-V-OPENAI.TOPIC-AI-LITIGATIONrelated_toTOPIC-FIRST-AMENDMENT.TOPIC-AI-LITIGATIONrelated_toTOPIC-DEVELOPER-SOFTWARE-LIABILITY.
Sources¶
SRC-THALER-V-PERLMUTTER-CADC: Thaler v. Perlmutter, 130 F.4th 1039 (D.C. Cir. 2025).SRC-USCO-AI-COPYRIGHTABILITY-REPORT-2025: Copyright Office AI copyrightability report, Part 2 (2025).SRC-THOMSON-REUTERS-V-ROSS-2025: Thomson Reuters v. Ross Intelligence (D. Del. 2025).SRC-BARTZ-V-ANTHROPIC-FAIR-USE: Bartz fair-use order (2025).SRC-BARTZ-V-ANTHROPIC-FINAL-APPROVAL: Bartz settlement final approval (2026).SRC-KADREY-V-META-FAIR-USE: Kadrey v. Meta (2025).SRC-NYT-V-OPENAI-MTD-2025: Times MTD opinion (2025).SRC-NYT-V-OPENAI-12C-2026: Times Rule 12(c) order (2026).SRC-SPYDER-V-MEMENTUM-COMPLAINT: Spyder complaint (2025).SRC-GARCIA-V-CHARACTER-MTD-2025: Garcia MTD order (2025).SRC-UMG-V-SUNO-COMPLAINT: UMG v. Suno complaint (2024).SRC-UMG-V-UDIO-COMPLAINT: UMG v. Udio complaint (2024).SRC-CONCORD-V-ANTHROPIC-MTD-2025: Concord MTD denial (2025).SRC-GEMA-V-OPENAI-LG-MUENCHEN-2025: GEMA v. OpenAI (2025).
Research Debt¶
- Add Walters v. OpenAI (defamation / ChatGPT hallucinations) with a primary docket and any opinion.
- Add Getty Images v. Stability AI (UK and U.S.) from judgments.
- Add EU AI Act public-enforcement matters only with official documents.
- Verify Garcia docket for any 2026 settlement or judgment before stating the case is resolved.
- Defamation, biometric-privacy, and employment AI cases are not yet in this inventory.
Document metadata¶
- Last verified:
2026-08-22
-
SRC-THALER-V-PERLMUTTER-CADC— 130 F.4th 1039 (D.C. Cir. 2025). ↩ -
SRC-USCO-AI-COPYRIGHTABILITY-REPORT-2025— Copyright Office Part 2 report (Jan. 2025). ↩ -
SRC-THOMSON-REUTERS-V-ROSS-2025— D. Del. Feb. 11, 2025. ↩ -
SRC-BARTZ-V-ANTHROPIC-FAIR-USE— N.D. Cal. June 23, 2025. ↩↩ -
SRC-BARTZ-V-ANTHROPIC-FINAL-APPROVAL— N.D. Cal. July 20, 2026. ↩ -
SRC-KADREY-V-META-FAIR-USE— 788 F. Supp. 3d 1026 (N.D. Cal. 2025). ↩↩ -
SRC-NYT-V-OPENAI-MTD-2025— S.D.N.Y. Apr. 4, 2025. ↩ -
SRC-NYT-V-OPENAI-12C-2026— S.D.N.Y. Aug. 6, 2026. ↩ -
SRC-SPYDER-V-MEMENTUM-COMPLAINT— N.D. Cal. Nov. 26, 2025. Pending; allegations. ↩ -
SRC-GARCIA-V-CHARACTER-MTD-2025— M.D. Fla. May 20, 2025. Pleading-stage order. ↩↩↩ -
SRC-UMG-V-SUNO-COMPLAINT— D. Mass. June 24, 2024. Pending. ↩ -
SRC-UMG-V-UDIO-COMPLAINT— S.D.N.Y. June 2024. Pending. ↩ -
SRC-CONCORD-V-ANTHROPIC-MTD-2025— N.D. Cal. Oct. 6, 2025. ↩ -
SRC-GEMA-V-OPENAI-LG-MUENCHEN-2025— LG München I Nov. 11, 2025. Not final. ↩