Anthropic has secured final approval for a landmark $1.5bn settlement with authors and book publishers who sued the artificial intelligence company over copyright infringement.

Judge William Alsup of the United States District Court for the Northern District of California approved the class-action settlement on Monday, clearing the way for payments to begin, according to Reuters.

The settlement provides for a payout of $3,000 per work across an estimated 500,000 copyrighted works, with the funds to be shared among authors and publishers holding the rights to the materials.

The settlement is believed to be the largest in the history of U.S. copyright law, although many authors and creators have questioned whether it represents a victory for copyright holders.

That is because the court had previously ruled in Anthropic’s favour on the central issue of whether training an AI model on copyrighted text constitutes fair use.

Alsup ruled that using copyrighted books to train an AI model could qualify as fair use, a decision widely viewed as significant for the rapidly expanding artificial intelligence industry.

However, the judge ruled that the way Anthropic obtained some of the books was unlawful.

The company built its training library from books it purchased and scanned, as well as millions of books downloaded from pirate websites, including Library Genesis and Pirate Library Mirror.

While the court found that the use of copyrighted text for AI training could constitute fair use, it determined that obtaining books through piracy was a separate legal issue that could proceed to trial.

Anthropic subsequently agreed to the settlement, avoiding a trial and potential damages that could have been awarded by a jury.

The final approval brings the case to a close but does not settle the broader copyright debate facing the AI industry.

Because Alsup’s ruling was issued by a single district court and Anthropic settled before an appeal, the decision will not establish binding precedent for other courts.