AI startup Anthropic has agreed to establish a $1.5 billion fund to compensate authors whose works were used without permission to train its models. The move follows a lawsuit brought by three writers—Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who accused the company of illegally ingesting their copyrighted books.
Court filings revealed that Anthropic admitted to acquiring millions of books, both by digitizing physical copies and by downloading pirated texts from notorious sites such as Library Genesis and Pirate Library Mirror. While the court sided with Anthropic on some points ruling that book scanning counts as fair use and that large language models (LLMs) create “transformative works” not bound by copyright, the company was found to have broken the law by knowingly incorporating pirated materials into its training data.
Rather than continue litigation, Anthropic proposed a settlement: it would create a $1.5 billion fund, valuing each pirated book at $3,000, and destroy the illicit copies. According to the filing, this outcome represents the largest copyright recovery claim to succeed in the U.S. Counsel for the plaintiffs noted it could “set a precedent of AI companies paying for their use of alleged pirated websites.”
The implications go far beyond Anthropic. Other major AI players including OpenAI and Perplexity AI face similar lawsuits over how they sourced training data. This case may also shape Anthropic’s ongoing dispute with Reddit over the scraping of forum content. Judges rarely reject proposed settlements of this kind, suggesting approval is likely.
Industry analysts say the Anthropic deal is just the beginning. Gartner vice president Daryl Plummer predicted that AI companies will increasingly be forced to set aside large sums to deal with legal challenges. “We’re going to see a lot more of this,” he said, warning that AI firms will need to create “slush funds” to settle copyright and liability claims. He went further, predicting, “Death by AI claims will rise one thousand percent,” as more people rely on AI advice, sometimes with harmful consequences.
Anthropic has not publicly commented on the settlement, but the company is well-funded; it announced $13 billion in new financing last week enough to absorb the costs of the payout.
