The Future of AI and Copyright: A Growing Global Divide

The landscape of AI innovation is increasingly shaped by an escalating copyright debate. In the U.S., rightsholders are aggressively pursuing legal action against AI companies accused of using copyrighted material without authorization. Meanwhile, other countries are taking a more flexible approach, permitting AI models to train on vast datasets, including those sourced from pirate libraries. This emerging “copyright schism” could have far-reaching consequences for the future of AI development.

The AI Copyright Divide

This week, various rightsholder organizations submitted recommendations for the 2025 Special 301 Report, an annual review by the U.S. Trade Representative highlighting countries that fail to meet U.S. copyright protection standards. Among their chief concerns was the role of copyright in AI training.

China has been singled out for considering a text and data mining (TDM) exception for AI development, a policy already implemented in Japan. This move has raised alarms among rightsholders and American tech leaders alike, who fear it may disadvantage U.S. companies facing stricter copyright enforcement.

The Battle Between Tech Giants and Copyright Holders

In the United States, AI training is not explicitly exempt from copyright restrictions. Consequently, major tech firms such as Meta, OpenAI, and Google are facing lawsuits for allegedly using unauthorized data sources, including pirate libraries, to train their large language models (LLMs).

Rightsholders argue that these repositories serve as a goldmine for AI, offering unrestricted access to vast amounts of unlicensed content. The legal debate now centers on whether this practice constitutes copyright infringement or falls under “fair use.”

These lawsuits are expected to take years to resolve. In the meantime, pirate libraries such as Z-Library, LibGen, and Anna’s Archive remain off-limits to U.S. AI companies. However, in countries with more relaxed copyright laws, the situation is markedly different, potentially creating an AI development gap between nations.

AI Companies and the Use of Shadow Libraries

One company that has drawn attention recently is DeepSeek, a Chinese AI firm that has released a highly efficient AI model. DeepSeek’s innovation challenges U.S. dominance in AI development by offering advanced capabilities at a fraction of the cost.

While DeepSeek’s recent publications are less transparent about their data sources, earlier studies explicitly referenced reliance on Anna’s Archive. One research paper published in March stated: “We cleaned 860K English and 180K Chinese e-books from Anna’s Archive.”

This practice is not isolated. Anna’s Archive has confirmed that multiple AI teams, including those affiliated with large U.S. and Chinese firms, have sought high-speed access to its datasets. The archive often collaborates with AI developers in exchange for financial contributions or data-sharing agreements. While most U.S. companies are cautious due to legal risks, many international teams operate with fewer constraints.

The Temptation of Unlicensed Data

For AI developers, shadow libraries represent an attractive, nearly irresistible source of knowledge. These massive, freely available datasets are akin to a “forbidden fruit”—immensely valuable yet fraught with legal peril.

In the U.S., AI companies risk lawsuits and hefty fines if caught using unauthorized data. This restrictive environment could place American AI development at a disadvantage, preventing access to the wealth of information available to international competitors.

Conversely, companies in countries with lenient copyright policies can freely train their models on whatever data they can access. This disparity could accelerate AI advancements outside the U.S., shifting the global balance of technological power.

The AI Copyright Dilemma

This divide raises critical questions about the intersection of copyright law and technological progress. Should all nations adopt stringent copyright policies to create a level playing field? Or should the West consider loosening its restrictions to remain competitive?

Rightsholders insist that global AI regulations must be strengthened to ensure fair compensation for copyrighted works. In contrast, proponents of open access argue that unrestricted knowledge sharing is essential for technological progress.

Anna’s Archive, for example, advocates for a radical shift in copyright policy, stating that “archiving and distributing books should be made fully legal” to maintain Western competitiveness in AI development.