Civil Groups Ask FTC to Probe AI Firms' Book Destruction

More than a dozen consumer advocacy and civil society organizations are calling on the Federal Trade Commission to investigate whether leading AI companies are engaging in anticompetitive practices by purchasing books, scanning them for training data, and then destroying the physical copies.

The groups — including Demand Progress Education Fund, the Consumer Federation of America, and the Institute for Local Self-Reliance — shared their letter with Axios before its public release. They argue that the practice goes beyond copyright concerns and may represent an unfair method of competition that could entrench dominant players.

Why it matters

This marks a potential shift in how regulators might approach AI training data disputes. Rather than focusing solely on intellectual property rights, the FTC could examine whether major AI firms are systematically eliminating resources that smaller competitors need to build rival models. If rare or out-of-print books are destroyed in the process, the competitive harm could be permanent.

The practice under scrutiny

According to court filings reported by the Washington Post in January, Anthropic spent millions of dollars acquiring physical books, removing their spines, and scanning the pages to train its Claude AI model. Google, Microsoft, and OpenAI have faced similar copyright lawsuits related to their training data practices, the Post reported.

The advocacy groups don't object to using books for AI training in principle. Instead, they're focused specifically on the destruction of physical copies after scanning. They argue this "hoard-and-destroy" approach could create what they call an "insurmountable systemic moat" around incumbent AI companies.

Competition concerns

The letter frames the practice as potentially "starving the market" of critical source materials. The groups note that in some cases, AI companies may be acquiring and destroying the last remaining copies of rare books, permanently removing them from circulation.

"We urge the FTC to view this practice not in isolation but rather as the latest escalation in a documented pattern of anticompetitive conduct," the groups write. They argue the practice could function as a structural mechanism to raise costs for rival companies and deny startups access to essential training materials.

The organizations are asking the FTC to act before AI companies can build lasting dominance through this approach.

Regulatory context

The request comes as the FTC under the current administration has sought to balance support for American business with scrutiny of Big Tech market power. Whether the agency will take up the investigation remains to be seen, but the letter represents a novel framing of AI training data issues through an antitrust lens rather than purely as a copyright matter.

The details were first reported by Axios.