AI copyright rulings favor tech giants, but may not be as ominous for publishers as they seem

Two recent copyright rulings favor AI companies, potentially making it easier for them to claim "fair use" over their use of copyrighted materials. However, the outcomes were mixed and instructive, pointing to potential avenues for litigants. The rulings do not necessarily close the door on copyright holders, and content creators should consult a lawyer specializing in copyright law for guidance.

Two recent copyright rulings in the Northern District of California have favored AI companies, potentially making it easier for them to claim "fair use" over their use of copyrighted materials. However, the outcomes were mixed and instructive, pointing to potential avenues for litigants. The rulings do not necessarily close the door on copyright holders, and content creators should consult a lawyer specializing in copyright law for guidance.

In Bartz v. Anthropic [1], District Judge William Alsup held that Anthropic engaged in fair use when it used purchased, copyrighted books for one-to-one destructive digitization and for training specific AI models. The judge ruled that the use of these works for training Claude, Anthropic's large language model, was transformative and favored fair use. However, Judge Alsup placed dispositive weight on the original manner of acquisition, rejecting the fair use defense for any use of pirated works. This decision underscores the importance of lawfully acquiring training data.

In Kadrey v. Meta [2], Judge Vince Chhabria granted summary judgment to Meta, finding its use of copyrighted books for training its "Llama" LLMs highly transformative. Judge Chhabria emphasized the fourth factor, market impact, and criticized the use of the "schoolchildren analogy." He found insufficient evidence of market harm to support the plaintiffs' case, leading to a victory for Meta. However, Judge Chhabria emphasized the narrowness of his holding and asserted that market dilution could often cause plaintiffs to win the fair use question overall.

These rulings contrast with the earlier District of Delaware ruling in Thomson Reuters v. ROSS [3], where the court found that training on copied material is direct infringement, and AI training does not necessarily mean fair use. This case highlights the need for AI companies to seek licenses or ensure lawful acquisition of training data.

The mixed outcomes of these cases suggest that copyright holders should focus on demonstrating concrete market harm, including indirect market substitution, when challenging AI training as infringement. AI developers must ensure all training data is lawfully acquired, and companies using third-party AI tools should seek strong indemnification and verify the provenance of training data to mitigate infringement risk.

References:
[1] https://www.ropesgray.com/en/insights/alerts/2025/07/a-tale-of-three-cases-how-fair-use-is-playing-out-in-ai-copyright-lawsuits
[2] https://www.ropesgray.com/en/insights/alerts/2025/07/a-tale-of-three-cases-how-fair-use-is-playing-out-in-ai-copyright-lawsuits
[3] https://www.ropesgray.com/en/insights/alerts/2025/07/a-tale-of-three-cases-how-fair-use-is-playing-out-in-ai-copyright-lawsuits