A group of writers, including Theranos whistleblower and author John Carreyrou, has filed a new lawsuit against six major AI companies, accusing them of training their models on pirated copies of their books. The lawsuit targets Anthropic, Google, OpenAI, Meta, xAI, and Perplexity, alleging that the companies engaged in copyright infringement by using stolen books to train their large language models (LLMs). This is not the first time authors have taken action against AI companies for similar reasons; a previous class-action suit against Anthropic resulted in a settlement of $1.5 billion, but some authors were dissatisfied with the outcome.

According to the new lawsuit, the plaintiffs argue that the proposed Anthropic settlement was unfair and did not hold the AI companies accountable for their actions. "The proposed settlement seems to serve the AI companies, not creators," said a spokesperson for the plaintiffs. "LLM companies should not be able to so easily extinguish thousands upon thousands of high-value claims at bargain-basement rates, eliding what should be the true cost of their actions."

The lawsuit highlights the complex issue of copyright infringement in the context of AI development. LLMs are trained on vast amounts of text data, which can include pirated copies of books. The companies argue that they are not liable for the actions of their predecessors, who may have obtained the books through illicit means. However, the plaintiffs argue that the companies are still responsible for the copyright infringement, as they benefit financially from the use of the stolen books.

This is not the first time authors have raised concerns about AI companies using pirated copies of their work. In 2022, a group of authors filed a class-action suit against Anthropic, alleging that the company had trained its LLM on pirated copies of their books. The case was settled out of court, with Anthropic agreeing to pay $1.5 billion to the affected authors.

The lawsuit also raises questions about the implications of AI development for society. As LLMs become increasingly prevalent, the issue of copyright infringement is likely to become more pressing. "The use of pirated books to train LLMs is a symptom of a larger problem," said a technology expert. "We need to rethink our approach to intellectual property in the age of AI."

The current status of the lawsuit is unclear, but it is likely to be closely watched by the tech industry and the literary community. The plaintiffs are seeking damages and a court ruling that would hold the AI companies accountable for their actions. If successful, the lawsuit could have significant implications for the development of AI and the way that companies approach intellectual property.

In a statement, a spokesperson for Anthropic said that the company was "reviewing the lawsuit and will respond accordingly." The other companies targeted in the lawsuit have not commented on the matter.