US Civic Groups Urge FTC Investigation into Allegations of Book Destruction by Big Tech

Buying books to scan and then burning them. Eighteen U.S. civic groups have called on the Federal Trade Commission (FTC) to investigate unfair competition practices, citing Big Tech AI firms' method of purchasing books to scan and subsequently destroying the originals in the process of developing cutting-edge AI technology. In a joint letter to the FTC on Tuesday (local time), organizations including the Demand Progress Education Fund and the Consumer Federation of America (CFA) argued that the practice of buying books in bulk to scan for AI training data and then discarding the originals constitutes 'input foreclosure,' preventing competitors from accessing the same materials.

The groups criticized the practice in their letter, stating that by permanently destroying large quantities of books, AI companies are blocking public access to these materials, arguing that this 'ensures only the wealthiest incumbent companies can build high-quality AI models.' They further claimed that by eliminating original copies of works, these firms are creating a future where specific entities exclusively possess humanity's records.

It has been revealed that some Big Tech companies have indeed acquired AI training data in this manner. Anthropic is known to have built its AI training data through the so-called 'Project Panama,' which involved buying millions of books, slicing off their spines to scan them, and then shredding them. Amazon is also reported to have trained its AI models in a similar manner.

The British daily The Guardian reported instances of bulk purchases of used books that appeared to be machine-selected from used bookstores in North America, Europe, and Australia. Tech outlet 404 Media reported that after attaching tracking devices to 1,000 books from a bookstore that received a large order, they confirmed that the books were delivered to an Amazon book scanning facility in Las Vegas, USA.