Anthropic reaches first US settlement in AI copyright dispute

AI startup Anthropic has resolved a class-action lawsuit filed by US authors, marking the first settlement in a growing wave of copyright battles against artificial intelligence companies.

Settlement marks a first in AI copyright cases

Artificial intelligence company Anthropic disclosed in a court filing on Tuesday that it has settled a class action lawsuit brought by a group of US authors.

The case centred on allegations that the company used copyrighted books without permission to train its AI systems.

The agreement represents the first settlement among several high-profile copyright lawsuits filed against AI developers in the United States.

These suits have targeted leading companies such as OpenAI, Microsoft, and Meta, as creators seek compensation for the alleged misuse of their works in training data sets.

While details of Anthropic’s settlement were not released, plaintiffs’ attorney Justin Nelson said in a statement that the resolution is a “historic settlement” that will benefit all class members.

He noted that further details are expected to be shared in the coming weeks. An Anthropic spokesperson declined to comment on the terms of the agreement.

Case stemmed from use of millions of books

The lawsuit was originally filed by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson in 2023.

They accused Anthropic—backed by major investors Amazon and Alphabet—of unlawfully downloading as many as 7 million books from pirate websites.

The works were allegedly saved in a central database and then used, at least in part, to train Anthropic’s Claude AI assistant.

In a June ruling, US District Judge William Alsup determined that Anthropic’s use of copyrighted works for AI training fell within the boundaries of “fair use.”

However, he also ruled that the company violated authors’ rights by maintaining a central library of pirated books that was not necessarily limited to AI training purposes.

Had the case proceeded to trial, Anthropic could have faced billions of dollars in damages.

US copyright law allows for statutory penalties of up to $150,000 per infringed work if the infringement is deemed willful.

Broader implications for the AI industry

The settlement comes at a time when AI developers are facing growing scrutiny over their reliance on copyrighted material to train large language models and generative AI systems.

Authors, publishers, and news organisations have filed multiple lawsuits seeking to establish boundaries around how copyrighted works may be used in training datasets.

For Anthropic, the settlement helps the company avoid a potentially lengthy and costly trial that had been scheduled to begin in December.

It also allows the firm to move forward without the immediate risk of massive financial liabilities, though the broader industry continues to grapple with unresolved legal questions.

The case highlights the tension between rapid innovation in AI and longstanding intellectual property protections.

While courts have offered mixed interpretations of how copyright law applies to AI training, settlements such as this one may set precedents for how disputes are handled in the future.

As other lawsuits progress against OpenAI, Microsoft, and Meta, the Anthropic case is likely to serve as a reference point.

For now, the terms of the agreement remain confidential, but industry observers expect the outcome to influence ongoing negotiations and litigation across the sector.