AI Copyright Lawsuit: Anthropic Reaches Settlement in Bartz v. Anthropic

Anthropic, a leading artificial intelligence company, has reached a settlement in the class action lawsuit Bartz v. Anthropic, addressing claims of copyright infringement related to the training of its large language models (LLMs). The lawsuit, filed in August 2024 by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, alleged that Anthropic utilized copyrighted works obtained from “shadow libraries” like Library Genesis (LibGen) to train its Claude LLMs .

The Core of the Dispute

The case centered on whether Anthropic’s use of copyrighted material to train its AI models constituted fair use or infringement. A pivotal summary judgment in June 2025, issued by Judge William Alsup of the Northern District of California, initially offered a split decision . The court found that Anthropic’s use of lawfully acquired books for AI training was “quintessentially transformative” and thus protected under fair use principles. However, the use of pirated copies obtained from shadow libraries was deemed infringing .

Settlement Details and FSF Involvement

Rather than proceed to trial on damages – which could have reached potentially hundreds of billions of dollars due to statutory copyright damages of up to $150,000 per infringed function – Anthropic and the plaintiffs opted for a settlement. The agreement involves offering monetary compensation to copyright holders in lieu of potential damages.

The Free Software Foundation (FSF) was among those who received notice of the settlement . The FSF’s copyright-protected work, Sam Williams’s Free as in freedom: Richard Stallman’s crusade for free software, was identified as having been used as training input for Anthropic’s LLMs . Although published under a free license (GNU Free Documentation License), the FSF views the unauthorized downloading of copyrighted works as a concern.

The FSF’s Stance on AI Training and Freedom

The FSF advocates for greater transparency and user freedom in the development and deployment of LLMs. They urge Anthropic and other AI developers to share complete training inputs, models, configurations, and source code with users . The organization believes that providing LLMs “in freedom” is crucial for protecting computing freedom and fostering innovation.

Key Takeaways

• The Bartz v. Anthropic case represents a significant legal test for the application of copyright law to AI training.
• The court distinguished between the fair use of lawfully acquired materials and the infringement associated with pirated content.
• The settlement avoids a potentially massive damages award and sets a precedent for compensating copyright holders affected by AI training practices.
• The FSF emphasizes the importance of user freedom and transparency in the development of LLMs.