Introduction
Meta’s recent legal victory in a copyright lawsuit brought by a group of 13 authors, including notable figures such as Ta-Nehisi Coates and Sarah Silverman [3] [7], underscores significant implications for the use of copyrighted works in AI training. The case [1] [3] [4] [7] [9] [10], centered on the LLaMA AI tool, highlights the complexities of copyright law in the context of artificial intelligence and the doctrine of fair use.
Description
Meta’s legal battle revolved around whether its practice of using the authors’ books without permission to train its AI models violated US copyright law. US District Judge Vince Chhabria ruled in favor of Meta, determining that the company’s actions fell under the fair use doctrine [5] [9]. The judge found that the plaintiffs failed to present compelling legal arguments or sufficient evidence to demonstrate that Meta’s actions would harm the market for their works, emphasizing that their claims of market dilution were speculative and without merit.
While the ruling acknowledges that copying protected works without permission is generally illegal [8], it clarifies that the range of output generated by Meta’s AI was too limited to substantiate the authors’ claims of harm [5]. The court identified three potential avenues for plaintiffs to argue market harm: the regurgitation of copyrighted works [3], harm to licensing markets for AI training [3], and the generation of works that closely resemble the originals [3]. However, the judge concluded that Meta’s AI systems were unlikely to saturate the market with similar works [5], and the authors did not demonstrate any actual market threat or economic loss resulting from Meta’s actions [5].
Meta welcomed the decision [2] [7], emphasizing the importance of fair use in the development of AI technologies [2]. The company defended its position by asserting that US copyright law allows for unauthorized copying to create transformative works [7], contending that LLaMA’s outputs are fundamentally different from the original texts [7]. They maintained that there is no evidence of LLaMA being used as a substitute for the authors’ books and argued that their methods of using the works do not affect the nature and purpose of the use [7]. Initially, Meta sought to license books from publishers [3], planning to invest up to $100 million in licensing deals; however [3], after negotiations stalled [3], the company opted to use unlicensed materials [3].
Judge Chhabria’s perspective on the use of pirated datasets for training suggests that Meta and others in the AI sector could be seen as serial copyright infringers, indicating that other authors might pursue similar cases in his court [7]. While a previous ruling favored AI startup Anthropic, which was found to have engaged in “exceedingly transformative” use of copyrighted works qualifying as fair use [10], concerns arise when companies create new products by copying protected works [11], as this could harm the market for the original materials [11]. The distinction between a general-purpose language model and one designed to create works similar to those it was trained on is crucial [6]. While the latter could be seen as less transformative [6], the act of training any large language model (LLM) is likely to retain some level of transformative quality.
This highlights the complexities of fair use analysis in the context of AI and copyright law [6], particularly regarding market effects and the potential for creative displacement [6]. Judge Chhabria expressed skepticism about the legality of companies using copyright-protected materials in their AI systems without permission or compensation [11], suggesting that in many cases [11], this would likely be illegal [11]. The ruling also acknowledged the potential market impact of generative AI, noting that it could undermine traditional creative processes by producing content more efficiently [2]. In a related case [8], a federal judge ruled that while training AI on copyrighted books could be fair use [3], downloading pirated copies for permanent storage violated copyright law [3]. The rulings from both judges leave critical questions about AI training and copyright law unresolved [3], paving the way for future litigation [3]. Other high-profile cases involving technology companies [3], such as OpenAI and Microsoft, are still pending [3], with plaintiffs likely to adapt their strategies in light of these early decisions [3].
Conclusion
The ruling in favor of Meta has far-reaching implications for the AI industry and copyright law. It underscores the importance of the fair use doctrine in the development of AI technologies while highlighting the ongoing legal uncertainties surrounding the use of copyrighted materials in AI training. As the legal landscape continues to evolve, future litigation will likely further define the boundaries of fair use in the context of AI, impacting both technology companies and content creators.
References
[1] https://variety.com/2025/biz/news/judge-rejects-claim-meta-ai-training-copyright-1236441270/
[2] https://www.straitstimes.com/world/united-states/meta-fends-off-authors-us-copyright-lawsuit-over-ai
[3] https://www.publishersweekly.com/pw/by-topic/digital/copyright/article/98093-meta-wins-ai-copyright-case-but-judge-writes-roadmap-for-authors-revenge.html
[4] https://news.bloomberglaw.com/litigation/meta-beats-copyright-suit-from-authors-over-ai-training-on-books
[5] https://www.analyticsinsight.net/news/big-win-for-meta-ai-copyright-case-outcome-could-shape-industry-future
[6] https://chatgptiseatingtheworld.com/2025/06/25/meta-prevails-on-fair-use-in-ai-training-in-kadrey-v-meta-but-judge-chhabria-cautions-a-better-record-of-dilution-or-market-harm-could-prevail-in-other-cases/
[7] https://www.pbs.org/newshour/arts/judge-tosses-authors-ai-training-copyright-lawsuit-against-meta
[8] https://aicommission.org/2025/06/meta-wins-ai-copyright-case-but-judge-says-others-could-bring-lawsuits/
[9] https://techcrunch.com/2025/06/25/federal-judge-sides-with-meta-in-lawsuit-over-training-ai-models-on-copyrighted-books/
[10] https://www.businessinsider.com/meta-wins-ai-copyright-lawsuit-lawyers-fumbled-boies-schiller-flexner-2025-6
[11] https://www.upi.com/Top_News/US/2025/06/26/Meta-AI-Llama-copyright-court-Facebook-Instagram/4961750946228/
