Court Approves Record Anthropic Payout to Authors — $1.5 Billion

A federal court for the Northern District of California has given final approval to an agreement under which Anthropic will pay $1.5 billion to a group of authors and publishers who accused the AI lab of copyright infringement. This was reported by the Reuters news agency. Preliminary approval of the deal was granted last year by Judge William Alsup, who has since retired; the final signature was provided by Judge Araceli Martínez-Olguín.

The payout will amount to $3,000 per work — around 500,000 such works were counted in all, and the money will be distributed among the authors and publishers who hold the rights. The settlement is considered the largest in the history of American copyright law, yet many writers and creators are in no hurry to call it a victory.

Piracy — Yes, Training — No

The reason lies in exactly how the legal question was resolved. On the key point, Alsup sided with Anthropic: he ruled that training an AI model on copyright-protected texts falls under fair use. This decision was seen as a turning point for the entire industry.

The judge did not, however, condone the way the company obtained the books. Anthropic assembled its training library from two sources: some of the books were purchased and scanned (to which there is no objection), while others were downloaded from pirate resources such as Library Genesis and Pirate Library Mirror. It was this second method that Alsup found to be unlawful in itself, and he allowed that the question of piracy might go to a jury. Soon afterward, Anthropic agreed to a settlement to avoid a trial and the unpredictable size of any potential award.

There Will Be No Precedent

The final approval closes this particular case, but it does not settle the matter for the whole industry. Alsup's ruling is the decision of a single district court, and since Anthropic chose to settle, the case will not reach an appeals court and will not become a binding precedent.

Other judges remain free to reach their own conclusions based on their own facts — which is exactly what is happening. A whole string of lawsuits continues against Google, Meta, Midjourney and OpenAI over the legality of training AI on protected works. Just last week a group of publishers and authors, including Hachette, Cengage, Elsevier, the writer Scott Turow and the organization S.C.R.I.B.E., filed a class-action lawsuit against Google, accusing the company of using their works to train its Gemini platform.

Source: TechCrunch