FTC urged to investigate AI firms destroying books

A coalition comprising more than a dozen civil-society organisations has formally asked the US Federal Trade Commission to determine whether these actions violate antitrust laws. As detailed by Businessaajkal, the group's request focuses on whether acquiring and physically destroying books constitutes an unfair method of competition under Section 5 of the Federal Trade Commission Act. Axios first reported on the letter sent to the federal watchdog.

Prominent organisations leading the push include the Demand Progress Education Fund, the Consumer Federation of America, and the Institute for Local Self-Reliance. Rather than attempting to halt training methods outright, the coalition's complaint targets the economic fallout of converting physical books into private digital datasets while removing the original volumes from circulation entirely.

Media additions

The coalition warns that larger firms are deliberately focusing on older titles, particularly those published before 2022, to guarantee the material was authored by humans and properly edited. These actions threaten to remove key resources from the open market permanently.

"Unlike a standard data acquisition strategy, this hoard-and-destroy practice could serve as yet another structural mechanism to raise rival companies' costs and deny start-ups and fledgling competitors a key source material essential to competing in the AI marketplace."

Civil society coalition, letter to the FTC, via Mashable and Businessaajkal

By absorbing available stocks of rare, specialised, and out-of-print titles, dominant players risk leaving themselves with the sole surviving digital copies of important works. The coalition describes the strategy as an escalation designed to construct an insurmountable systemic moat around established operators, urging regulators to view the tactic not in isolation but as part of a documented pattern of anticompetitive conduct.

This regulatory push follows earlier disclosures regarding specific industry practices. Court records revealed that Anthropic acquired and destructively scanned millions of physical books under an internal initiative known as Project Panama to gather high-quality written text for training the AI models behind Claude. The process involved removing book spines for rapid page scanning before sending the physical remnants away for recycling.

Mashable noted that the trend of AI companies using and destroying physical copies of books to train agentic AI is spooking even devotees of artificial intelligence due to its resemblance to book burning. OpenAI CEO Sam Altman has previously described knowledge as a utility that could be sold to customers like electricity and water, a business proposition the coverage notes looks less wholesome when paired with the wholesale destruction of human knowledge repositories that can otherwise be acquired cheaply or loaned freely from local libraries.

Key Arguments and Concerns

• Market Exclusion: Large companies use significant financial resources to buy up books in bulk, locking smaller developers out of necessary training resources and raising rival companies' costs.
• Irreplaceable Loss: Rare, out-of-print, and specialised works are permanently removed from circulation and cannot be replaced once destroyed, threatening key source material.
• Antitrust Focus: The complaint shifts the debate from copyright infringement and author compensation to fair competition, market dominance, and the exclusion of fledgling competitors under Section 5 of the Federal Trade Commission Act.

The coalition's appeal introduces a competition-policy dimension to a debate that has largely centered on copyright, licensing, and whether technology companies should compensate authors and publishers for using their work. The Federal Trade Commission did not immediately announce whether it will open a formal investigation into the matter.