Imagine receiving a massive order for rare books, only to realize the buyer is an AI company scanning them to train its models. That's the reality for one independent bookseller, who initially dismissed the inquiry as spam. This incident highlights a growing trend where AI firms are consuming physical books in bulk, often leaving sellers with damaged or destroyed inventory.

The 'Spam' Order That Wasn't

A bookseller recently received an unusually large request for books. At first glance, it seemed like a scam or spam—perhaps a phishing attempt or a prank. But the request was genuine. The buyer was an AI company looking to scan the books to build datasets for training large language models.

This is not an isolated event. Across the industry, booksellers are reporting similar inquiries from AI companies. These firms often purchase entire collections, not to read or resell, but to digitize and use the content for machine learning. The result is that books are often cut, scanned, or otherwise destroyed in the process.

Why AI Companies Need Physical Books

AI models require vast amounts of text data to improve their understanding of language and knowledge. While much of this data comes from the internet, physical books offer a unique and high-quality source of information. They contain long-form, well-structured content that is often not available online.

By scanning books, AI companies can access copyrighted material that is not freely available on the web. This has led to a surge in demand for physical books, particularly in niche subjects and older publications. However, the process is not without controversy, as it raises questions about copyright and fair use.

The Scanning Process Is Destructive

To scan a book efficiently, many companies use automated systems that require the book to be taken apart. This often means cutting off the spine and feeding pages through a scanner. The result is that the original book is rendered unusable, which is a significant loss for collectors and libraries.

In the case of the bookseller, the large order would have meant destroying a significant portion of their inventory. This is why the initial request seemed so improbable—it's not common for someone to pay for books and then discard them.

Impact on Independent Booksellers

For independent booksellers, these bulk orders can be both a blessing and a curse. On one hand, they represent a significant sale, which is rare in a struggling industry. On the other hand, they can deplete valuable stock that took years to acquire.

Moreover, the ethical implications weigh heavily on many sellers. They are torn between making a profit and preserving literary heritage. Some have started to decline such orders or negotiate terms that allow for non-destructive scanning.

This trend is likely to continue as AI companies seek more data. Booksellers may need to adapt by setting clear policies about how their books are used.

Key Takeaways

  • AI companies are acquiring physical books in bulk for scanning to train models, often destroying the books in the process.
  • Booksellers are often caught off guard by these requests, which can seem like spam or scams.
  • The practice raises copyright and ethical concerns, as it involves reproducing copyrighted material without explicit permission.
  • Independent booksellers face a dilemma between profit and preservation.

As artificial intelligence continues to evolve, the demand for data will only grow. It remains to be seen how the publishing and bookselling industries will respond to this new challenge.