A group of prominent writers, led by investigative journalist and Bad Blood author John Carreyrou, has filed a new lawsuit accusing several major artificial intelligence companies of unlawfully using pirated books to train their AI models. The defendants named in the suit include Anthropic, Google, OpenAI, Meta, xAI, and Perplexity, marking a significant escalation in the legal battle between creative professionals and the fast-growing AI industry.
The case follows an earlier class action lawsuit against Anthropic that ended in a $1.5 billion settlement. Although that agreement offered eligible authors payouts of about $3,000 each, it failed to satisfy some writers. A court had ruled that while training AI systems on copyrighted books could be lawful, the act of illegally copying those books was not. The new plaintiffs argue that the settlement effectively shielded AI companies from meaningful accountability for allegedly acquiring training data through unlawful means.
In court filings, the authors contend that the previous settlement primarily benefited AI firms rather than creators, allowing companies to resolve thousands of potential claims at what they describe as deeply discounted costs.
They insist that AI developers should bear the full financial consequences of what they call deliberate and large-scale copyright infringement, especially as AI products continue to generate billions of dollars in revenue. The lawsuit seeks to reopen the debate over how AI models are trained and whether current legal remedies adequately protect authors’ rights.
