Anthropic settles AI copyright lawsuit brought by authors

Anthropic reached a settlement to pay USD 1.5 billion to authors in a landmark copyright lawsuit over its use of pirated books to train its AI model Claude, depriving them of income. What happened AI company A nthropic settled a class-action lawsuit brought by authors Andrea Bartz, Kirk Wallace Johnson and Charles Graeber, who accused the company of using unauthorised copies of approximately 500,000 books obtained from pirate libraries such as LibGen and PiLiMi (Pirate Library Mirror) to train its language model Claude. The settlement agreement requires Anthropic to pay USD 1.5 billion (USD 3,000 per book) and destroy the pirated content it used. It does not affect claims regarding AI outputs or future conduct. The lawsuit arose after a court ruling found Anthropic’s use of purchased books constituted fair use, but ruled that the use of pirated works was infringing and not protected by fair use, describing piracy of copyrighted works as "inherently, irredeemably infringing." The settlement, which requires the approval of US District Judge William Alsup, is the largest publicly reported copyright recovery in history. Why it happened Anthropic’s legal defense in the AI copyright lawsuit centered on the argument that its use of copyrighted materials to train its AI model Claude constituted "fair use" under U.S. copyright law. Specifically, Anthropic contended that digitising books it lawfully purchased and using those digital copies for machine learning was highly transformative and essential for competitive AI development. The company argued that its training approach was analogous to human reading and learning, turning the content into something fundamentally different rather than replicating the originals. However, Anthropic acknowledged it had also used over seven million digital copies of books obtained from pirate websites. What it means The settlement offers authors unprecedented compensation and recognition of harms from the unauthorised use of their works to train AI models, and sets a financial and legal precedent for handling copyrighted content. It signals that companies using third-party content to train their models must seriously consider licensing copyrighted materials - or face costly litigation. System 🤖 Claude 🔗 Developer: Anthropic Country: USA Sector: Media/entertainment/sports/arts Purpose: Multi-purpose Technology: Genera tive AI; Machine learning Issue: Accountability; Copyright; Transparency Regulation ⚖️ Copyright Act of 1976 Legal, regulatory 👩🏼‍⚖️ Bartz et al. v. Anthropic PBC

Date it happened

2022-01-01

Organisation involved

Anthropic

Product, system or model

Claude

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.