London's rare bookshops appear to have been targeted by AI companies seeking books to train their models. This follows court documents revealing that Anthropic, an AI developer, bought, scanned, and destroyed millions of physical books as part of a project to gather training data.
A London-based rare bookseller told City AM they received two unusually large enquiries for specialist books from anonymous email addresses, lacking names or company details. Both buyers requested additional photographs before placing significant orders, but neither replied when asked about their identity or the books' destination.
These enquiries mirror reports from across Europe, where antiquarian booksellers have described similar requests for thousands of obscure titles from anonymous buyers. 404 Media also reported a rise in bulk purchases by second-hand booksellers, suspected to be orchestrated by AI companies concealing their identities.
Court documents showed Anthropic's internal programme, 'Project Panama', was initiated after the company concluded books were "essential" for training advanced AI models. The process involved buying millions of second-hand books, slicing off their spines with industrial cutting machines, scanning every page, and then recycling the remainder.
A US federal judge later ruled that this scanning process was protected by fair use, as Anthropic legally purchased each book, converted it into a digital copy, and destroyed the original. The company had previously agreed a $1.5bn (£1.13bn) settlement over claims related to its earlier use of pirated digital books.