UK and Ireland Used Booksellers Report Unusual Bulk Orders Suspected to Be AI Data Collection

Secondhand booksellers in the UK and Ireland are reporting mysterious bulk orders they suspect come from AI companies seeking text to train models.

Secondhand booksellers across the UK and Ireland have begun reporting an unusual pattern: large, seemingly indiscriminate bulk orders that sellers suspect are being placed by or on behalf of artificial intelligence companies seeking raw text data for model training. The suspicion is not unfounded. Anthropic, the AI company behind the Claude assistant, was previously found to have spent millions of dollars purchasing physical books specifically to scan them as part of what was described as 'data acquisition' for training its AI systems. That revelation has put the broader bookselling trade on alert for similar activity from other firms. The pattern booksellers describe as 'strange' involves orders that do not follow the typical logic of a private collector or reseller — quantities and subject matter that suggest institutional rather than individual purchasing. Independent secondhand dealers, who often operate on thin margins, are now trying to determine whether the apparent boom in their sales represents a genuine windfall or a legally and ethically complicated relationship with the AI industry. The broader question of whether AI companies have the right to scan and train on copyrighted books without author or publisher consent is already the subject of active litigation in several countries, making the sourcing of physical copies a notable workaround some firms appear to be pursuing.

Why it matters

If AI companies are systematically purchasing secondhand books to circumvent digital copyright controls, it raises significant legal and ethical questions for authors, publishers, and the book trade. It also signals how far AI firms are willing to go — and spend — to acquire training data at scale.

What's next

Booksellers and publishing industry observers will be watching to see whether AI companies formally acknowledge such purchasing programs and whether existing copyright litigation addresses the physical-book scanning loophole.

Key facts

• Booksellers in the UK and Ireland are reporting bulk orders they describe as 'strange' and suspect originate from AI companies
• Anthropic was previously found to have spent millions of dollars on books specifically to scan them for AI training data
• The suspected purchases are described as institutional in scale and pattern, unlike typical collector or reseller orders
• Physical book purchases may represent an attempt to acquire copyrighted text outside the scope of current digital licensing agreements
• The practice is emerging against a backdrop of active copyright litigation targeting AI companies in multiple countries

Bias & framing notes

The Guardian frames the story around suspicion and potential wrongdoing by AI firms, anchored by the Anthropic precedent. The BBC headline takes a more neutral, even positive framing ('sales are booming'), though no body text was available to assess its full angle. Because only one source provided substantive reporting, the full picture — including any AI company responses or denials — may be missing.

NewsClear — neutral news & congressional tracking · Bill of the Week