AN FRANCISCO (AP) — A federal judge last week approved a $1.5 billion settlement between artificial intelligence company Anthropic and authors who allege nearly half a million books had been illegally pirated to train chatbots.

U.S. District Judge William Alsup issued the preliminary approval in San Francisco federal court after the two sides worked to address his concerns about the settlement, which will pay authors and publishers about $3,000 for each of the books covered by the agreement. It does not apply to future works.

The Association of American Publishers called the settlement a “major step in the right direction in holding AI developers accountable for reckless and unabashed infringement.”

San Francisco-based Anthropic said it is pleased with the preliminary approval.

Alsup had dealt the case a mixed ruling in June, finding that training AI chatbots on copyrighted books wasn’t illegal but that Anthropic wrongfully acquired millions of books through pirate websites to help improve its Claude chatbot.