The AI copyright lawsuits that started as individual actions are becoming coordinated industry campaigns. In February, fourteen major publishers, among them Conde Nast, Vox Media, The Atlantic, The Guardian, and Axel Springer, jointly filed suit against Cohere, the enterprise AI company. It is the largest media coalition in the AI copyright disputes to date, and it signals that publishers are done waiting for individual settlements to set precedent.
What the publishers are alleging
The core claim across most of these cases follows the same pattern: AI companies harvested copyrighted content from pirate libraries, including LibGen, Z-Library, and OceanofPDF, in addition to the open web, and used that material to train their models without licensing agreements or payment to rights holders. Publishers argue this isn't fair use. It's systematic reproduction that directly competes with the original works.
Individual authors have made the same argument at smaller scale. Six writers recently filed suit against Anthropic, OpenAI, Google, Meta, xAI, and Perplexity, arguing that the specific value those models produce comes substantially from having absorbed their work. They're seeking more than $3,000 per copyrighted title, according to Publishers Weekly.
The Anthropic settlement
The Bartz v. Anthropic case has produced the first major settlement in this wave of litigation. Anthropic agreed to pay $1.5 billion. Preliminary approval came in September 2025, and the final approval hearing is set for May 14, 2026. The Authors Guild outlines the structure: funds distribute equally across all eligible copyrighted titles after legal and administrative costs.
That structure is worth examining. A bestselling novel and an obscure self-published title receive identical distributions. Whether that's equitable is a separate argument from whether the overall sum is adequate. Many authors believe the settlement undervalues their work, and some are actively objecting to the payout formula.
What comes next
The outcome that will affect all current litigation is the Thomson Reuters v. Ross Intelligence appeal, scheduled to be heard in June. That case will determine whether using copyrighted material to train AI constitutes fair use. A ruling against fair use would accelerate the licensing negotiations that several AI companies have already quietly begun with publishers, and likely trigger another wave of settlements at much higher valuations.
The 14-publisher coalition against Cohere suggests media companies aren't waiting for that clarity. If individual settlements take years to finalize, filing collectively cuts legal costs, increases negotiating leverage, and creates pressure that a single publisher suing alone cannot generate.
Commentarii · 0