Anthropic's $1.5B copyright settlement approved
The approval settles a landmark class-action over training AI with copyrighted books and sets a potential precedent for fair use in AI training while leaving unresolved the broader legal questions of piracy and future damages.
At a glance
- A federal judge approved a $1.5 billion settlement between Anthropic and authors/publishers, distributing about $3,000 per work across roughly 500,000 works.
- The ruling, which approved the settlement, followed a finding that Anthropic’s use of copyrighted text could be fair use for training AI, though piracy via sites like Library Genesis was deemed illegal.
- The decision is not a binding precedent for the industry since it is a district court ruling, and the case will not be appealed to set broad precedent.
The story
TechCrunch reported that a federal judge gave final approval to Anthropic’s $1.5 billion class-action copyright settlement, ending a high-profile dispute over AI training on copyrighted books. The payout is approximately $3,000 per work across an estimated 500,000 works, to be shared among authors and publishers who hold rights.
The settlement follows a ruling by Judge William Alsup (now retired) that training an AI model on copyrighted text can count as fair use, a decision widely viewed as a watershed moment for the AI industry. The agreement acknowledged that Anthropic obtained books from two sources: purchased/scanned copies and pirate sites such as Library Genesis, with the piracy method deemed illegal on its own terms.
Although the settlement resolves the class action, TechCrunch notes that Alsup’s fair-use ruling was limited to a single district court and thus does not create binding precedent for other cases. The article also notes that more copyright lawsuits against AI firms continue in other courts.