The AI Copyright Litigation Landscape
AI training-data copyright cases became a defining legal issue of 2024-2026. Plaintiffs include news publishers, authors, music labels, image-rights holders, and visual-art creators; defendants include all major AI labs. Outcomes will shape AI economics for years. This page tracks active and resolved cases as of May 2026.
Key Findings
- More than 35 distinct AI training-data lawsuits were active in major Western jurisdictions in Q2 2026, plus dozens of smaller and international actions.
- The largest cases by potential damages are New York Times v OpenAI / Microsoft, the Authors Guild class action, Universal Music v Anthropic, and Getty Images v Stability AI.
- Major settlements and licensing deals are reshaping the landscape: Reddit-OpenAI, AP-OpenAI, FT-OpenAI, News Corp-OpenAI, Conde Nast-OpenAI, Time-OpenAI, plus Anthropic licensing deals.
- The fair-use defence has produced mixed early rulings; some courts have been receptive to transformative-use arguments, others sceptical of training-data appropriation at scale.
- Training-data licensing markets emerged in 2024-2025 with ProRata, ScalePost, TollBit, and direct publisher-to-AI-lab deals.
Major Active Cases (Q2 2026)
| Case | Plaintiff | Defendant | Filed | Status |
|---|---|---|---|---|
| NYT v OpenAI / Microsoft | The New York Times Company | OpenAI, Microsoft | Dec 2023 | Active in SDNY; discovery underway |
| Authors Guild class action | Authors Guild + named authors | OpenAI | 2023 | Active; consolidated proceedings |
| Universal v Anthropic | Universal Music + others | Anthropic | 2023 | Active; lyrics-output cases |
| Getty v Stability AI | Getty Images | Stability AI | 2023 | UK and US; trial-stage |
| Concord et al. v Anthropic | Music publishers | Anthropic | 2023 | Active |
| Sarah Silverman et al. | Authors | OpenAI / Meta | 2023 | Active |
| Andersen v Stability / Midjourney / DeviantArt | Artists | Multiple AI image cos | 2023 | Active |
| Daily News et al. v OpenAI / Microsoft | News publishers | OpenAI, Microsoft | 2024 | Active |
| Center for Investigative Reporting | CIR | OpenAI, Microsoft | 2024 | Active |
| Major studios v Midjourney | Disney, Universal | Midjourney | 2025 | Active |
Notable Settlements and Licensing Deals (through Q2 2026)
| Counterparties | Approximate value | Year | Type |
|---|---|---|---|
| Reddit / OpenAI | ~$60M/yr | 2024 | Multi-year licensing |
| Reddit / Google | ~$60M/yr | 2024 | Multi-year licensing |
| News Corp / OpenAI | ~$250M over 5 yrs | 2024 | Licensing |
| FT / OpenAI | Not disclosed | 2024 | Licensing + product partnership |
| AP / OpenAI | Not disclosed | 2023 | Licensing |
| Conde Nast / OpenAI | Not disclosed | 2024 | Licensing |
| Time / OpenAI | Not disclosed | 2024 | Multi-year |
| Vox Media / OpenAI | Not disclosed | 2024 | Multi-year |
| Hearst / OpenAI | Not disclosed | 2024 | Multi-year |
| Anthropic / various publishers | Not disclosed | 2024-2025 | Multiple |
Legal Developments in 2026
- Multiple summary judgement decisions on fair-use defences shape doctrine; rulings have been mixed
- UK Information Commissioner's Office and EU Data Protection authorities issued guidance on training-data lawfulness
- EU AI Act's Article 53 training-data summary requirements created compliance obligations independent of copyright
- Music industry settlements set precedent for music-label / AI-lab licensing structures
- Class-action consolidation produced larger but slower-moving litigation
Training Data Licensing Markets
Three vendor categories emerged to facilitate training-data licensing:
- Aggregator marketplaces: ProRata, ScalePost, TollBit aggregate publishers and license to AI labs collectively
- Direct publisher-to-lab deals: News Corp / OpenAI, Conde Nast / OpenAI, etc.
- Pay-per-crawl infrastructure: Cloudflare Pay-Per-Crawl, similar primitives that monetise individual fetches
Implications for AI Lab Economics
Training-data licensing has shifted from optional to operationally significant:
- Estimated 2025 licensing spend across major AI labs: $1.5-3 billion (highly uncertain)
- Per-training-run licensing budget for frontier models: $50-200 million (estimated)
- This is small relative to compute costs but growing, and creates competitive moats favouring well-capitalised labs
Brand Visibility Implications
Training-data lawsuits and licensing deals are extensively covered by legal, technology, and media industry journalism, generating substantial inbound links. Brands selling content licensing infrastructure, AI legal services, training-data marketplaces, content authentication, or content-licensing legal counsel face high AI-mediated discovery surface as media companies, AI labs, and counsel query AI assistants for related vendor recommendations.
Methodology
Case tracking from court dockets (PACER, UK Royal Courts of Justice), specialist trackers, plaintiff and defendant filings, and press reporting. Settlement values often non-public; figures are estimates from press disclosures. Updated quarterly as cases progress.
How Presenc AI Helps
Presenc AI tracks brand-mention rates inside AI assistant queries about AI legal services, content licensing, training-data marketplaces, and AI copyright matters. For vendors operating in this space, this is the operational visibility into a discovery surface tightly coupled to media and legal industry attention.