Court approves compensation of US$ 1.5 billion from Anthropic to authors for copyright

The largest copyright agreement ever recorded, worth US$1.5 billion, was formally approved this Monday, July 20, 2026, ending a complex dispute between the artificial intelligence company Anthropic and thousands of authors. The decision was handed down by federal judge Araceli Martínez-Olguín, setting a significant precedent for the future use of AI-protected content and consolidating the largest compensation ever awarded in copyright actions.
Details of the largest copyright compensation ever approved
The sentence ends a class action that became a judicial landmark. The process began after Anthropic was accused of using literary works to train its artificial intelligence models without due authorization. Although the court initially considered AI training to be “fair use”, piracy of works was singled out as a likely copyright infringement. Authors initially challenged the settlement, arguing that attorneys’ fees were excessive and payments to writers were insufficient.
The main decisions of judge Araceli Martínez-Olguín
Judge Araceli Martínez-Olguín rejected the objections to the agreement, considering them without merit. In her order, she emphasized the high participation rate in the lawsuit: approximately 95% of class members had been notified, and approximately 91% of affected authors and publishers had already filed their claims. Only 350 members chose not to participate, while another 54 attempted to object or submitted opt-out requests after the deadline.
The judge noted that the high turnout indicated that the majority of plaintiffs received notification of the distribution plan in a timely manner, considered it “fair” and supported the agreement. She also highlighted that the estimated payments of $3,000 per work were “four times the minimum statutory damages.”
Some crucial points of the court decision include:
- Amount approval:The total value of the agreement of US$1.5 billion was fully approved.
- Cut in lawyer fees:The lawyers’ initial request for 20% of the settlement ($300 million) was reduced to 12.5% (approximately $187 million) before the final decision. The judge, however, considered 12.5% still too high and cut the fees to less than 7% of the settlement fund, resulting in about $101 million.
- Reduction of awards to main authors:The three plaintiffs who represented the class in the litigation had their service awards reduced from $50,000 to $15,000 because the higher amount was deemed “unreasonable.”
Author challenges and delayed exit attempts
During the process, some authors attempted to oppose the agreement or request the deletion of their works after the deadline, seeking higher damages in separate lawsuits. The main arguments revolved around the claim that they did not receive timely notice, the exclusion period was too short, and that the Copyright Act would allow for potentially greater statutory damages.
Judge Martínez-Olguín confirmed that 350 class members were able to exclude themselves from the settlement, but dozens of objections and at least nine direct requests for exclusion after the March 30 deadline were denied. The court accepted only two late exclusion requests, both from co-authors who demonstrated “excusable negligence” by failing to meet the deadline – one of them due to health problems and a language barrier in Mexico.
One of the most notable cases was that of Donald Passman, an entertainment lawyer and author, who submitted his application in June, three months after the deadline. Martínez-Olguín called the request “inexplicably delayed” as records showed that notices were delivered to his address. Anthropic defended the refusal, arguing that allowing Passman’s late exclusion would set precedents and delay the distribution of the funds.
The impact and next steps for artificial intelligence companies
The $1.5 billion deal has broad implications for the artificial intelligence industry. The lawsuit’s lead plaintiffs said the outcome “brings us closer to real accountability for Anthropic and puts all AI companies on notice that they cannot break the law or ignore creators’ rights.”
Anthropic, in turn, expressed satisfaction with the resolution of the case. In a statement, its deputy general counsel, Aparna Sridhar, expressed delight with the decision that established that its AI training falls under fair use, in addition to high buy-in from rights holders.
Martínez-Olguín highlighted that the agreement offers important non-monetary benefits to authors. They avoid costly, time-consuming and complex litigation, which could result in loss without any financial recovery. Additionally, the agreement requires Anthropic to destroy all used works and allows for future lawsuits if the company misuses them after the process is complete. There is also the possibility of redistributing any remaining funds after all valid claims have been paid, which may result in payments exceeding the initially anticipated US$3,000 per work.






