In the long arc of human civilization, libraries have stood as shared inheritance — the accumulated thought of generations held in common. Now, consumer advocacy groups are asking the Federal Trade Commission to examine whether major AI companies are quietly dismantling that inheritance, buying books in bulk, extracting their knowledge to train artificial minds, and then destroying the physical originals — some of them the last copies in existence. Filed with the FTC on Friday by more than a dozen organizations, the complaint frames this alleged 'hoard-and-destroy' practice not merely as cultu
Consumer groups urge FTC probe into AI firms' 'hoard-and-destroy' book practices
Only the wealthiest incumbents can build high-quality AI models
Mark
Why does it matter if a company destroys books after scanning them? They bought them legally, didn't they?
Mimi
They did buy them legally, but that's not the whole story. Once you destroy the last copy of a book, no one else can ever access it—not researchers, not other AI companies, not the public. It's a permanent loss of a resource that can't be replaced.
Mark
So the concern is that this gives one company an unfair advantage?
Mimi
Exactly. If you're wealthy enough to buy millions of books, scan them, and then destroy them so no one else can use them, you've essentially locked up the training data. Smaller competitors can't access the same material, even if they have the money to buy books—because the books are gone.
Mark
But couldn't other companies just buy their own books?
Mimi
Not if the books have been destroyed. And some of these are the last surviving copies. Once they're gone, they're gone forever. The groups are arguing this creates a monopoly not just on AI quality, but on access to humanity's written heritage.
Mark
What would the FTC actually investigate?
Mimi
They'd want to know the scale—how many books are being destroyed, how many are irreplaceable, and whether the practice violates antitrust law. The legal question is whether destroying books after using them constitutes an unfair method of competition.
Mark
Has anyone won a case on this yet?
Mimi
Not yet. A court ruled that Anthropic could legally use purchased books to train AI. But that didn't address whether destroying them afterward crosses an antitrust line. That's what the FTC might explore.
O Pulso
- More than a dozen consumer advocacy groups have filed an urgent FTC complaint alleging that AI companies are systematically buying, scanning, and destroying millions of books — including irreplaceable last-surviving copies — to lock competitors out of the same training data.
- The stakes are existential in a quiet way: once these books are gone, they are gone forever, and the knowledge they contain becomes the exclusive property of whichever corporation was wealthy enough to digitize them first.
- A 2025 court ruling already found that Anthropic's use of legally purchased books to train its Claude model was permissible under copyright law, but that decision left the antitrust dimension — the deliberate destruction to foreclose competition — entirely unaddressed.
- Reporting by 404 Media this week added Amazon to the picture, alleging the company is running a parallel operation: bulk-buying books, scanning them for AI tools, and discarding the originals.
- The FTC now faces a defining choice — investigate the scale and legality of these practices under Section 5 of the FTC Act, or allow a quiet privatization of the written record to continue unchallenged.
In the long arc of human civilization, libraries have stood as shared inheritance — the accumulated thought of generations held in common. Now, consumer advocacy groups are asking the Federal Trade Commission to examine whether major AI companies are quietly dismantling that inheritance, buying books in bulk, extracting their knowledge to train artificial minds, and then destroying the physical originals — some of them the last copies in existence. Filed with the FTC on Friday by more than a dozen organizations, the complaint frames this alleged 'hoard-and-destroy' practice not merely as cultural loss, but as a calculated consolidation of power — one that could leave only the wealthiest AI incumbents holding the keys to humanity's written heritage.
A coalition of more than a dozen consumer advocacy groups, including the Demand Progress Education Fund and the Consumer Federation of America, filed a complaint Friday with the Federal Trade Commission asking it to investigate what they call a 'hoard-and-destroy' strategy among major AI companies. The allegation: that these companies are purchasing books in bulk, digitizing them to train AI language models, and then destroying the physical copies — deliberately preventing competitors and the public from accessing the same material.
What gives the complaint its particular weight is the nature of what may be lost. Many of the books allegedly being discarded are among the last surviving copies of their works. The advocacy groups warn that this creates a landscape where only the wealthiest AI developers can build high-quality models, while smaller competitors and the general public are left without access to the source material those models consumed. In their words, AI companies risk '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 complaint has real precedent behind it. Writer Andrea Bartz and others sued Anthropic in 2024, accusing the company of acquiring, scanning, and discarding millions of print books. A court ruled in 2025 that using legally purchased books to train AI was permissible under copyright law — but that ruling said nothing about whether the subsequent destruction of those books raises antitrust concerns, which is precisely the angle the advocacy groups are now pressing. Separately, an investigation by 404 Media this week found Amazon engaged in similar practices, buying books in bulk, scanning them for AI tools, and destroying them afterward.
Neither Anthropic nor Amazon responded to requests for comment. If the FTC chooses to investigate, it would mark a significant moment in the collision between AI development and the public's claim to its own cultural inheritance.
A coalition of consumer advocacy groups has asked the Federal Trade Commission to investigate whether major artificial intelligence companies are engaging in what critics call a "hoard-and-destroy" strategy—buying books in bulk, digitizing them to train their AI systems, and then destroying the physical copies to prevent competitors and the public from accessing the same material.
The complaint, filed Friday with FTC Chairman Andrew Ferguson and Commissioner Mark Meador, comes from more than a dozen organizations including the Demand Progress Education Fund, the Consumer Federation of America, and the Institute for Local Self-Reliance. They argue that this practice is not only destructive but fundamentally anticompetitive, potentially violating Section 5 of the FTC Act, which prohibits unfair or deceptive business practices in commerce.
What makes the allegation particularly troubling is the irreplaceable nature of some of the destroyed material. According to the groups' letter, many of the books being discarded are among the last surviving copies of their original works. Once destroyed, they cannot be recovered. The advocates worry that this creates a scenario where only the wealthiest AI companies—those with enough capital to buy and digitize vast libraries—can build high-quality language models. Everyone else, including smaller AI developers and the general public, loses access to the source material that made those models possible.
The groups framed the concern in stark terms: "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." They are asking the FTC to determine how widespread the practice is and how many of the destroyed books represent the last known copies of their works.
The complaint is not theoretical. In August 2024, writer Andrea Bartz and others sued AI developer Anthropic, accusing the company of acquiring, scanning, and discarding millions of print books. A judge ruled in 2025 that Anthropic's use of legally purchased books to train its AI model, Claude, was permissible under copyright law. That ruling, however, did not address whether the subsequent destruction of those books raises separate antitrust concerns—the angle the advocacy groups are now pushing the FTC to examine.
Recent reporting has also implicated Amazon in similar practices. An investigation by digital publisher 404 Media this week found that Amazon is buying books in bulk, scanning them for its AI tools, and then destroying them. Neither Anthropic nor Amazon responded immediately to requests for comment on the allegations.
The investigation, if the FTC chooses to pursue it, would mark a significant moment in the ongoing tension between AI development and public access to information. The question at stake is whether companies should be allowed to use books as training material and then eliminate them from circulation, or whether doing so crosses a line into anticompetitive behavior that harms both the public interest and fair competition in the AI industry itself.
Citações Notáveis
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.— Advocacy groups in their FTC complaint
Quer a matéria completa?
Leia o original em CBS News ↗
Technology