John Carreyrou Sues Major AI Firms Over Alleged Copyrighted Book Use in AI Training

Investigative journalist John Carreyrou, widely known for exposing fraud at Silicon Valley blood-testing startup Theranos and for authoring the bestselling book Bad Blood, has filed a new copyright infringement lawsuit against several leading artificial intelligence companies. The lawsuit, filed Monday in a California federal court, targets Elon Musk’s xAI, Anthropic, Google, OpenAI, Meta Platforms, and Perplexity, accusing them of using copyrighted books without permission to train their large language models.

Carreyrou, a New York Times reporter, filed the complaint alongside five other writers. The plaintiffs allege that the AI companies unlawfully copied and ingested their books into the datasets used to build and refine AI chatbots and other generative AI systems. According to the lawsuit, this practice amounts to large-scale piracy that violates U.S. copyright law and deprives authors of compensation for the use of their work.

The case adds to a growing wave of legal challenges brought by authors, publishers, and other rights holders against technology companies over AI training practices. Notably, this lawsuit is the first copyright case to name xAI, Elon Musk’s artificial intelligence venture, as a defendant. Representatives for the companies involved did not immediately respond to requests for comment.

Unlike other high-profile AI copyright cases, the plaintiffs are not seeking class-action status. The writers argue that class actions tend to benefit defendants by enabling them to resolve thousands of claims through a single settlement. The complaint states that such arrangements allow AI companies to “extinguish” numerous high-value copyright claims at relatively low cost.

The lawsuit references a recent settlement involving Anthropic, which agreed in August to pay $1.5 billion to resolve claims from a class of authors who alleged the company pirated millions of books. According to the new filing, authors in that settlement will receive only about 2% of the maximum statutory damages allowed under the Copyright Act for each infringed work.

The complaint was filed by attorneys from Freedman Normand Friedland, including Kyle Roche, who was previously profiled by Carreyrou in a 2023 New York Times article. Carreyrou has publicly criticized prior settlements, calling the alleged theft of books to build AI systems the industry’s “original sin” and arguing that previous agreements failed to adequately protect authors’ rights.