Meta has secured a pivotal victory as U.S. District Judge Vince Chhabria granted a summary judgment in favour of Meta in Kadrey v. Meta – the lawsuit brought by 13 authors, including Sarah Silverman and Ta‑Nehisi Coates – who alleged unauthorised use of their books to train Meta’s Llama AI. The judge concluded that the plaintiffs failed to show sufficient market harm, a key element in copyright infringement claims. Importantly, though the decision favoured Meta, Chhabria emphasised that it does not establish a sweeping legal precedent allowing AI firms to freely train models on copyrighted material.
This ruling follows closely on the heels of another court victory for Anthropic, where Judge William Alsup similarly found AI training on books to be fair use, but flagged unlawful acquisition through pirated sources. Chhabria’s ruling echoed Alsup’s caution: transformative use alone is insufficient without demonstrable market impact. He noted that the authors had “made the wrong arguments” and left no room for proving market dilution.
For global tech firms and AI developers, this marks a double-edged moment. On one hand, courts are recognising transformative training as fair use under current doctrine. On the other, judges are warning developers that blanket defence won’t suffice – future cases with better evidence of financial harm could overturn this trend. AI companies must now tread carefully: strengthen documentation of transformative use, avoid questionable source materials, and consider proactive licensing of copyrighted content.
Looking forward, the ruling is unlikely to halt legal momentum in copyright circles. Chhabria’s decision limits itself to the 13 plaintiffs and the specific evidence presented. As other creators refine claims around market displacement, courts may face tougher tests, and potentially set more restrictive boundaries for AI training practices.
In sum, while Meta celebrates a strategic win in June 2025, the broader copyright debate over generative AI remains fraught. Technology leaders must now balance innovation with legal diligence, anticipating a future where fair use is narrowly defined, and market impact evidence becomes vital in defending AI training practices.
