AI Companies Accused of Hoarding and Destroying Books for Training Models

A coalition of public interest and consumer advocacy groups has called on the Federal Trade Commission (FTC) to investigate leading artificial intelligence developers. The organizations accuse these AI companies of a "hoard-and-destroy" practice: acquiring books in bulk, digitizing their content to train large language models, and then destroying the physical copies.

In a letter to FTC leadership, groups including the Demand Progress Education Fund and the Consumer Federation of America argue that this practice is not only "destructive" but also anticompetitive. They contend that destroying books, sometimes the last surviving copies of original works, could severely limit public and other AI companies' access to essential source material. This could create a future where only the wealthiest AI developers can access and control humanity's written works.

The advocacy groups are asking the FTC to assess the scale of this alleged practice and determine how many destroyed books might be unique or last remaining copies. Their concerns are amplified by a 2024 copyright lawsuit against AI developer Anthropic, which was accused of acquiring, scanning, and discarding millions of print books. Although a judge previously ruled that Anthropic's use of legally purchased books for AI training did not violate copyright, the "hoard-and-destroy" allegations persist.

Furthermore, a recent investigation by digital publisher 404 Media reported that Amazon is also engaged in buying books in bulk, scanning them for AI training, and subsequently destroying them. Both Anthropic and Amazon did not immediately respond to requests for comment on these reports. The groups argue that such actions could constitute an unfair method of competition under the FTC Act.

“

Through their practice of permanently destroying books en masse and thus removing those nonrenewable resources from broader access, AI companies are engineering a future where only the wealthiest incumbents can build high-quality AI models and operate as the sole holders of humanity's written works, after having destroyed the originals to get there.