As part of its ongoing monitoring of international developments, the Hellenic Copyright Organization (HCO) highlights a landmark decision by the U.S. federal courts that sets an unprecedented precedent regarding the relationship between Artificial Intelligence and copyright. U.S. Federal Judge Araceli Martínez-Olguín has granted final approval to a US$1.5 billion (approximately €1.3 billion) settlement between Anthropic and thousands of authors and publishers. This is the largest financial settlement in the history of copyright worldwide.
Key points of the decision:
- The infringement: Anthropic was held liable for the large-scale downloading and storage of millions of copyrighted works from illegal “pirate” digital libraries for the purpose of training its AI chatbot, Claude.
- Legal distinction: The Court ruled that training AI models using lawfully acquired books may qualify as fair use, whereas the use of pirated material constitutes a clear copyright infringement.
- Compensation: Authors and rights holders will receive US$3,000 per work, covering approximately 500,000 books, while 91% of eligible claimants have already submitted their claims.
- Deletion of data: Anthropic is required to permanently delete all files obtained from the illegal sources.
This landmark settlement is expected to have a significant impact on the pending copyright lawsuits against other major AI companies, including OpenAI, Google, and Meta.
For further information, please refer to Reuters and the official Anthropic Copyright Settlement