Anthropic can move ahead with a $1.5 billion settlement over copyrighted books after a federal judge gave final approval, TechCrunch reports. The deal covers a class action brought by authors and publishers and is expected to pay $3,000 per work across roughly 500,000 works.
The case matters because it split two issues that are often blended together. Earlier, Judge William Alsup ruled that training an AI model on copyrighted text can qualify as fair use. But he also found that Anthropic illegally downloaded and stored books from pirate sites, including Library Genesis and Pirate Library Mirror.
Anthropic settled after the piracy issue was allowed to proceed toward trial. That means the final approval closes this specific dispute, but it does not create binding appeals-court precedent for the wider AI industry.
Other copyright lawsuits against Google, Meta, Midjourney, OpenAI, and others are still moving through courts. The settlement gives rightsholders compensation in one case, while leaving the larger question of how copyright law applies to AI training unresolved.