Judge: Meta Wins but Transformative Element of Fair Use Doesn't Outweigh Market Impact
- Josh Waterston
- Jun 26
- 2 min read
"Every battle is won before it is ever fought." - Sun Tzu.
Meta just won an early skirmish in the battle over whether AI training is fair use - but AI companies may end up losing overall if they focus on only the transformative element of copyright law's fair use test without addressing the market impact.
Yesterday, in Kadrey v. Meta, Judge Chhabria granted Meta summary judgment, and said that even if training AI models on copyrighted works is transformative - "As a factual matter, there’s no disputing that" - the market effects arguments are likely to win out. AI platforms might not diminish the market for highly recognized works. "But you can bet that the market for lesser-known biographies of Lyndon B. Johnson will be affected. And this, in turn, will diminish the incentive to write biographies in the future."
Judge Chhabria also noted Judge Alsup's ruling from two days prior in Bartz v. Anthropic, in which Judge Alsup held that AI training was transformative and discounted the market effects factor, saying that Judge Alsup's "inapt analogy is not a basis for blowing off the most important factor in the fair use analysis."
The court also dismissed the argument that upholding copyright will destroy AI companies:
"The technology is certainly groundbreaking. But the suggestion that adverse copyright rulings would stop this technology in its tracks is ridiculous. These products are expected to generate billions, even trillions, of dollars for the companies that are developing them. If using copyrighted works to train the models is as necessary as the companies say, they will figure out a way to compensate copyright holders for it."
Judge Chhabria held that the plaintiffs' specific arguments weren't convincing, but he provided a roadmap for future plaintiffs to follow:
"No matter how transformative LLM training may be, it’s hard to imagine that it can be fair use to use copyrighted books to develop a tool to make billions or trillions of dollars while enabling the creation of a potentially endless stream of competing works that could significantly harm the market for those books."
As the judge pointed out, news sites may have strong arguments against fair use by AI platforms that summarize their articles and siphon off their web traffic and ad/subscription revenue. However, non-profits using AI to perform medical research may have a stronger fair use argument.
In summary: the fair use analysis remains unchanged. Just because judges have found that training is "transformative" doesn't mean that Meta and Anthropic have won the battle. There's a long road ahead. If you're a business looking to train an AI model, you still need to respect copyright.
If you're a creator or business with questions about how to navigate this rocky road, please reach out - I'd be glad to discuss.
Orders:
