**A group of well-known writers, including investigative journalist John Carreyrou, has filed a new lawsuit accusing major AI companies of training their models on pirated books, arguing that earlier settlements failed to address the core issue of how that material was obtained.**

![Authors Sue AI Firms Over Alleged Use of Pirated Books ](https://www.stanventures.com/news/wp-content/uploads/2025/12/ChatGPT-Image-Dec-24-2025-03_15_48-PM-300x200.png)

A new copyright lawsuit is adding pressure to some of the biggest names in artificial intelligence. This time, the plaintiffs include award-winning authors who say their books were copied and used without permission to train large language models.

The [lawsuit](https://fingfx.thomsonreuters.com/gfx/legaldocs/zdvxjkawlvx/AUTHORS%20AI%20COPYRIGHT%20LAWSUIT%20complaint.pdf) names Anthropic, Google, OpenAI, Meta, xAI, and Perplexity as defendants. 

At the center of the case is the claim that these companies trained their AI systems using pirated copies of copyrighted books.

Among the authors bringing the case is[John Carreyrou](https://en.wikipedia.org/wiki/John_Carreyrou), best known for reporting on the Theranos scandal and writing Bad Blood. 

The complaint was filed this week and builds on earlier legal action taken by other writers against Anthropic.

## Why Authors Are Back in Court

The lawsuit comes after a federal judge ruled in a previous case that [training AI models on copyrighted material](http://stanventures.com/news/adobe-faces-new-lawsuit-over-ai-training-data-6356/) could be allowed under fair use. However, the same ruling also made an important distinction. While training itself may be lawful, illegally copying or downloading books is not.

That earlier case ended in a $1.5 billion settlement with Anthropic. Eligible authors could receive payments of around $3,000. Some writers accepted the deal, but others felt it missed the point.

The authors behind the new lawsuit argue that the settlement closed thousands of claims without meaningfully addressing how the books were acquired. In their view, the agreement benefited AI companies more than the people whose work was used.

The complaint states that the settlement appeared designed to protect companies from further liability rather than hold them responsible for using unauthorized copies of books.

## What the Authors Are Arguing

The lawsuit states that AI companies have generated enormous commercial value from models trained on copyrighted books. 

The authors say that resolving thousands of claims for modest payouts ignores the scale of the alleged infringement.

One section of the complaint criticizes what it calls bargain-rate resolutions that eliminate claims without reflecting the real market value of the content involved. The authors argue that allowing such settlements sets a dangerous precedent for creative industries.

Their position is direct. Technology companies should not be able to rely on unauthorized copies of books, build profitable systems, and then settle claims in a way that limits accountability.

## Why This Matters Beyond This Case

This legal challenge highlights a growing tension between creators and AI developers. Courts are still figuring out how copyright law applies to machine learning, and each ruling shapes how future models are built.

If the authors succeed, AI companies may face stronger expectations to license books and other creative works directly. That could increase costs and change how training data is collected.

If the companies prevail, it may reinforce a system where disputes are resolved through settlements that limit broader accountability, even when the source material was improperly obtained.

## What Authors Should Pay Attention To

This case highlights the importance of staying aware of how digital copies of books move across the internet. Publishing contracts, licensing terms, and distribution rights play a larger role than many authors realize when disputes like this arise.

The lawsuit is also being watched closely for how courts handle one key distinction: whether training AI models can be treated separately from how the source material was obtained. How judges answer that question could influence future legal action and negotiations across the publishing industry.

## What Comes Next

The case is expected to move into early court motions that will determine whether it stands apart from earlier cases and can move forward. One area likely to draw close scrutiny is how responsibility is defined when copyrighted material is copied before it is used to train AI systems.

With [AI tools](http://stanventures.com/blog/ai-content-writing-tools/) becoming more widely used, disputes like this are set to remain a major part of the broader debate over how technology companies and creators share value and responsibility.

## Key Takeaways

- Authors have filed a new lawsuit over alleged use of pirated books
- Six major AI companies are named as defendants
- The case follows dissatisfaction with a prior Anthropic settlement
- Authors say settlements avoid accountability for unauthorized copying
The outcome could influence future AI training and licensing practices