John Carreyrou Sues Google, xAI, OpenAI Over AI Training

John Carreyrou, the New York Times investigative reporter and author of Bad Blood, joined five other writers in filing a federal copyright lawsuit on December 22, 2025 in U.S. District Court in California, alleging that multiple technology companies used their articles and books without authorization to train large language models that power chatbots and other AI services.

The complaint accuses developers of "pirating" copyrighted work by ingesting reporters' articles and authors' books into training datasets and then using those models to power products without seeking permission or compensating rights holders. The filing names Google, xAI and OpenAI among the defendants. Other versions of the complaint and related summaries identify additional companies including Anthropic, Meta Platforms and Perplexity, reflecting a broader list of AI developers that plaintiffs contend relied on copyrighted material.

The suit joins a growing wave of litigation by news organizations, authors and publishers challenging the practices of AI firms. Plaintiffs seek to hold major developers accountable for the inclusion of copyrighted books and news articles in model training, arguing that the unauthorized use of expressive, news and literary content has deprived creators of control over their work and potential revenue streams.

The case follows earlier high profile actions. In 2023, The New York Times sued OpenAI and Microsoft, alleging their systems were trained on millions of Times articles without compensation. The Times later filed a separate suit against Perplexity, asserting that the company repeatedly retrieved and republished substantial Times content through its AI powered answer engine. Publishers and news outlets worldwide have pursued similar claims, and other plaintiffs have asked courts for remedies ranging from damages to orders barring further use of copyrighted content in training.

The complaint filed by Carreyrou and his co plaintiffs focuses on core copyright theories that have yet to be definitively resolved by the courts in the context of large language models. Central legal questions include whether ingesting full articles and books into training datasets constitutes a public performance or reproduction, and whether downstream uses of those trained models can be characterized as fair use when they generate text informed by copyrighted sources.

Legal analysts say the arrival of this suit is notable for explicitly listing xAI, the company founded by Elon Musk, among defendants. That inclusion marks one of the first reported instances of xAI being named in a copyright suit tied to model training. The filings do not, in the versions available, specify the full list of allegedly used works, the identities of the other five plaintiffs, or the precise remedies sought.

Defendants had not filed a public response in the court docket summaries provided at the time of filing. The immediate procedural schedule, including any motions or hearings, was not specified in the complaint excerpts circulated with the filing notice.

Beyond the courtroom, the litigation underscores a broader industry tension. Publishers and creators argue that widespread ingestion of their reporting and books threatens both the ad supported model of news and authorship income, while tech companies maintain that access to diverse textual sources is integral to building useful AI systems. How judges treat those competing interests in the coming months could reshape content licensing, model design and the economics of journalism and publishing.