Meta, the parent company of Facebook, has won a copyright lawsuit filed by authors including Sarah Silverman and Ta-Nehisi Coates. The authors accused Meta of using pirated copies of their books without permission to train its AI model, Llama. However, US District Judge Vince Chhabria ruled that the plaintiffs failed to provide adequate evidence that Meta's AI system harmed the market for their works.

While the ruling was in Meta’s favor, the judge clarified that this case does not confirm that training AI with copyrighted content is legal. Instead, it highlights that the plaintiffs didn’t present the right legal arguments.

Chhabria emphasized that in many circumstances, using copyrighted materials without permission to train AI would be unlawful. This nuanced stance contrasts with another San Francisco judge's ruling earlier this week in favor of AI firm Anthropic, which deemed such usage fair use.

The fair use doctrine is a crucial defense for AI companies, allowing limited use of copyrighted materials without explicit permission. Meta welcomed the decision, describing fair use as vital to building transformative AI systems.

Meanwhile, the authors’ legal team criticized the judgment, citing an “undisputed record” of Meta’s large-scale use of copyrighted works.

The broader copyright battle between AI companies and creators continues to intensify. Companies like OpenAI, Microsoft, and Anthropic face multiple lawsuits from writers, journalists, and publishers over the use of copyrighted materials in AI training.

Judge Chhabria warned of a future where AI-generated content could flood the market, undermining the value of original human-created works and disincentivizing creativity. Despite the legal win, the debate over AI’s use of protected content is far from over.