A high-profile legal battle over artificial intelligence training is taking shape in California, as a veteran investigative journalist from the New York Times has sued several of the world’s most influential AI companies. The case adds fresh momentum to a growing debate about whether tech firms crossed legal boundaries when using copyrighted books to build powerful chatbots now used by millions.
The lawsuit was filed by John Carreyrou, a Pulitzer Prize-winning reporter for The New York Times and author of the bestselling book “Bad Blood,” which exposed fraud at the now-defunct blood-testing company Theranos.
Carreyrou joined five other writers in accusing a group of AI developers of using their books without permission to train large language models, the systems that generate human-like responses in chatbots.
Among the defendants named in the complaint are OpenAI, Google, Meta Platforms, Anthropic, Perplexity and xAI, the artificial intelligence company founded by Elon Musk.
According to the lawsuit, these companies allegedly copied and ingested copyrighted works to improve their AI products, without authorization or compensation to the authors.
The case was filed in federal court in California and marks the first time xAI has been formally named as a defendant in a copyright lawsuit related to AI training. While lawsuits against AI companies have multiplied over the past two years, most have focused on a smaller group of developers. This new filing broadens the legal spotlight to include nearly every major player shaping the future of generative AI.
In their complaint, the writers argue that AI firms have benefited enormously from the creative labor of authors while avoiding the costs typically associated with licensing or purchasing books.
They claim this practice has allowed companies to accelerate development while undermining the economic rights protected by U.S. copyright law.
Unlike other cases currently moving through the courts, the plaintiffs in this lawsuit deliberately chose not to pursue a class action. They argue that class actions often benefit defendants by enabling them to resolve thousands of claims through a single settlement that may not reflect the true value of each infringed work. Instead, the writers are seeking individual accountability and potentially higher damages per book.
The lawsuit also references a major settlement reached earlier this year involving Anthropic, which agreed to pay $1.5 billion to resolve claims from authors who said their books were used without permission.
However, Carreyrou and the other plaintiffs contend that authors in that case will ultimately receive only a small fraction of the maximum damages allowed under copyright law, raising questions about whether such settlements truly deter future violations.
Carreyrou has been particularly vocal about the issue. In court proceedings linked to the Anthropic case, he described the alleged use of pirated books as the company’s “original sin,” arguing that financial settlements alone do not fully address the harm done to writers or the broader creative ecosystem.
Responses from the companies named in the lawsuit have been limited so far. A spokesperson for Perplexity said the company does not index books, while representatives for the other defendants did not immediately comment. As the case moves forward, it is likely to intensify pressure on AI developers to clarify how their models are trained and whether current practices comply with copyright law.
The outcome could have wide-ranging implications, not only for authors and publishers, but also for the future development of artificial intelligence. At stake is a fundamental question: whether innovation built on vast amounts of human-created content can continue without clearer rules, consent and compensation.
