June 25th: Meta Beats Authors’ Copyright Suit—But Judge Warns Future AI Training May Still Violate Law

It appears that big tech is on a bit of a win streak as a relates to recent copyright litigation. Meta, on the immediate heels of the favourable Anthropic judgement, just gained a favourable ruling in its own fair use case; although it certainly wasn’t an overwhelming victory, despite asserting otherwise.

A US District Court judge granted summary judgment to Meta in Kadrey v. Meta, a closely watched case in which 13 authors—including Sarah Silverman and Ta-Nehisi Coates—alleged that the company’s Llama models were trained on pirated copies of their books without permission. Judge Vince Chhabria ruled the plaintiffs failed to prove concrete market harm, a key element of copyright infringement, and therefore dismissed the claim.

Why this matters

• Narrow victory, broad implications. While Meta escapes liability here, Chhabria’s opinion stresses that copying copyrighted works for AI training “will be illegal” in many circumstances unless creators are compensated—signaling that future plaintiffs who produce solid evidence of market damage could still prevail.

• Contrast with Anthropic ruling. The decision lands just days after another California judge sided with Anthropic on fair-use grounds. Together, the cases highlight diverging judicial views on how “transformative” AI training must be and revive debate over what constitutes economic harm to rightsholders.

• Industry reaction. Meta hailed the ruling as a win for fair use, asserting that open-source AI “powers transformative innovations.” Authors’ advocates called the outcome “limited,” noting the door remains open for broader claims and potential damages once evidence of market dilution is properly presented.

What’s next

Legal experts expect a wave of refined lawsuits targeting whether AI companies used pirated datasets and how output might substitute for original works. The judgment effectively invites future claims that quantify lost sales, licensing revenue, or audience engagement—metrics creative industries are now rushing to document. For developers, the message is clear: negotiate licenses or be ready to prove your training data causes no economic harm.

Final thoughts

It is clear that copyright and fair use will remain highly litigated topics, as content creators fight to prevent Big Tech from gaining unfettered access to their IP, in order to train the numerous models coming to market. Where do you stand? Do you support Big Tech or the Content Creators in this battle for content “ownership”?