Amazon's Rare Book Operation Raises Questions

Artificial intelligence users may soon gain access to a new feature: digital scans of rare and antique books. The technology could open up accessibility to works that are otherwise difficult to find, but it may also threaten years of work by translators and collectors.

The story broke a few weeks ago, when 404 Media reported that a shipment of rare books was tracked to an Amazon warehouse in Nevada. There, the books are scanned, translated, summarized, and printed word for word into an AI digital database, intended to help with recall and context for new prompts on older books. The move follows the lawsuit filed by authors against the AI company Anthropic, which was accused of also scanning books for AI use and destroying them. 

The purpose of this operation, according to 404 Media’s Emanuel Maiberg in an interview with NPR, is to avoid what it describes as "model collapse",  a phenomenon in which AI trained on other AI output produces varied and incoherent answers. The solution, in this case, is for the system to consume human-generated text to produce more human-like responses.

What makes the operation especially striking is the company behind it: a firm known for a dinosaur logo holding a book in its hands,  almost as if depicting the dinosaur as a kind of shredder. That imagery raises further questions about intentionality and the core values of those involved, as destructive scanning has been on the minds of the public even though Amazon heavily denies it. 

Yet, after the analysis is complete and the books are scanned, sorted, and stripped of their words and themes, they are destroyed in their entirety. These books could’ve been donated and passed on to new readers, but their lifespan is cut short to make the process efficient and to give companies a legal loophole if ever tried in court, as they’re not pirating.

 Rare books have never been printed in large quantities, and they serve as important records of outdated products and materials no longer in use. With so many being intentionally destroyed, many worry that their contents will be lost in translation, without the human interpretation and value that give these items meaning.

The practice raises a broader question: where is the line between helpful access and the dilution of authors' messages and translations built up over the years? These are books that many readers cannot obtain anywhere else. With some suggesting the company is trying to monopolize rare books and capitalize on niche genres and categories that only it can offer access to. 

The extreme measures make one wonder what else Amazon will do to turn users to its platform, and its platform alone. For many, the feature doesn't seem like a necessity. Rare books are already, in many ways, hidden behind museum exhibits and online archive paywalls. In an age when hundreds of books are banned each year, and access is being shortened, there’s been a push on readers to rely on AI systems and processes. Interest in literature as a whole has been steadily declining, but could AI scans create more exposure for rare books in the long run?  We’ll just have to wait and see.