A coalition of more than a dozen public interest and consumer watchdog organizations is pressing the Federal Trade Commission to scrutinize major artificial intelligence companies over allegations that they have purchased, digitized and then destroyed books to build large language models — a practice the groups say is both “destructive” and potentially harmful to competition.
In a Friday letter sent to FTC Chairman Andrew Ferguson and Commissioner Mark Meador, organizations including the Demand Progress Education Fund, the Consumer Federation of America and the Institute for Local Self-Reliance claimed that AI firms are acquiring large quantities of books, scanning their contents for use in AI training and disposing of the physical volumes afterward. The letter alleges that, in some instances, the books being destroyed may include some of the final existing copies of original works.
Critics describe the alleged strategy as a “hoard-and-destroy” approach, warning that it could limit access to important source material for the public, researchers and rival artificial intelligence developers. The groups also argue that destroying books after digitizing them may amount to an unfair method of competition under Section 5 of the FTC Act, which prohibits unfair or deceptive practices in commerce.
“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,” the letter states.
The organizations are asking the FTC to investigate how widespread the alleged practice is and to determine how many destroyed books may have been the last known copies of the works they contained.
The letter points to an August 2024 copyright case filed by author Andrea Bartz and other writers against AI company Anthropic, accusing the developer of obtaining, scanning and discarding millions of printed books. In 2025, a judge found that Anthropic’s use of lawfully purchased books to train Claude, its AI model, did not violate copyright law.
Anthropic did not immediately reply to a request for comment.
The complaint also follows a report this week from digital publisher 404 Media, which said Amazon has been purchasing books in large batches, scanning them for AI training and then destroying the copies. Amazon did not immediately respond to a request for comment regarding the report.
Alain Sherter
