A federal judge in California has granted final approval to a $1.5 billion settlement between artificial intelligence company Anthropic and a group of authors who accused the company of improperly using copyrighted books to develop its Claude chatbot, marking the largest copyright settlement in U.S. history and the first major resolution of a wave of lawsuits targeting AI model training practices.
According to Reuters, U.S. District Judge Araceli Martínez-Olguín approved the agreement on July 20 in San Francisco, rejecting objections from some class members who argued that the settlement amount was insufficient. The case stems from allegations that Anthropic copied millions of books, including pirated works, to build datasets used in training its large language models.
The litigation, known as Bartz v. Anthropic, has become one of the most closely watched legal battles in the emerging field of generative AI. Reuters reported that the settlement covers claims related to more than seven million books that plaintiffs alleged were collected and stored by Anthropic, with compensation already claimed by more than 91% of eligible authors and publishers.
Anthropic, which is backed by Amazon and Alphabet, said the settlement was reached after earlier court rulings established that using books to train AI systems could qualify as fair use under U.S. copyright law. The company has maintained that the agreement addresses historical data acquisition practices and noted that it has since shifted toward obtaining training materials through licensed and lawful channels.
The dispute gained particular significance following a June 2025 ruling by then-U.S. District Judge William Alsup, who found that the act of training AI models on copyrighted books constituted transformative use protected under the fair use doctrine. However, the court also concluded that retaining and storing pirated copies of books could still expose the company to copyright liability. Reuters reported that potential statutory damages in the case could have reached hundreds of billions of dollars had the matter proceeded to trial.
Judge Martínez-Olguín also approved attorneys’ fees totaling more than $101 million, significantly below the amount originally requested by plaintiffs’ counsel, according to court filings and subsequent reporting.
Related: One Government, Two Voices: Recent Anthropic vs. Pentagon Developments Add to the Record Before the D.C. Circuit
The settlement arrives amid an expanding legal campaign against major AI developers over the sourcing of training data. Authors, publishers, artists, and news organizations have filed dozens of lawsuits against companies including OpenAI, Meta Platforms, Google, Microsoft-backed entities, and xAI, alleging that copyrighted material was used without authorization to train generative AI systems.
The broader dispute has increasingly drawn attention from policymakers and competition regulators, who are examining whether control over massive datasets and computing infrastructure could reinforce market concentration in the rapidly growing AI industry. Regulators in the United States, United Kingdom, and European Union have separately scrutinized partnerships between leading AI developers and large technology companies, including investments by Amazon, Microsoft, and Alphabet in AI startups.
Competition authorities have expressed concern that access to proprietary data, cloud infrastructure, and distribution channels may create barriers for smaller developers, potentially concentrating power among a handful of dominant firms. While the Anthropic settlement itself is a copyright matter rather than an antitrust case, legal experts have argued that disputes over training data licensing could influence the competitive structure of the AI sector by determining which companies can secure lawful access to large-scale content repositories.
The case may also accelerate the emergence of a commercial market for AI training licenses. Several publishers and media companies have already entered licensing agreements with AI developers, creating new revenue streams while attempting to establish clearer legal frameworks for the use of copyrighted material in model development.
Reuters reported that some authors and publishers opted out of the Anthropic settlement and continue to pursue separate claims against the company. Additional lawsuits against other AI developers also remain pending, leaving several unresolved questions regarding fair use, compensation standards, and the future regulation of training data practices in the generative AI industry.
Source: Reuters
