The latest disclosure in the artificial intelligence industry is arguably the most concerning, and it’s taking place on our bookshelves.
According to recent reporting, AI firms are buying books in bulk, removing their bindings, scanning them, and then destroying the originals to create data used to train large language models. It’s a process called destructive scanning.
Video footage obtained by News.com.au from ABTec Solutions, a Canadian digitisation company, showed the process in action, with the bindings of countless books systematically sliced off – rendering them virtually useless to a reader – in preparation for scanning.
Perhaps most worrying of the trend, booksellers across Europe and the US say they’re receiving unusual bulk orders for rare, specialist and out-of-print titles, raising concerns that some of these works could disappear from bookshelves all together.
A small book seller spoke to 404 Media, giving an insight into the scale at which this is happening, saying he went from selling around 20 books per week to hundreds – particularly hard-to-find and out-of-print books.
“It benefits me financially as well as by clearing out old inventory that is otherwise unlikely to sell,” he said. “I’ve been well suited for these sales with inventory from overseas and foreign language books. On the other hand, I don’t like the end-use, and I don’t like that uncommon books are being pulped.”
According to Dutch news site BNR, similar requests are being received across Europe. One antique bookseller told the publication he had received an email request from a company in Singapore requesting copies of 3000 English-language titles across various categories. Although the email claimed it was for “a new project focused on collecting books in multiple languages”, many suspect AI companies are driving the charge. That’s because the practice has emerged as the AI industry searches for new sources of training data, with much of the internet now exhausted. The appeal of print books published before 2022, then, is that they’re free from AI-generated content. But while the race to innovate speeds up, we need to ask ourselves what we’re losing in the process.
ISBNdb is one company promoting the practice. It purports to have the “world’s largest book database”, and has started offering books in bulk for the purposes of AI. The website reads: “ISBNdb gives AI teams access to one of the largest structured book metadata databases available” spanning “111 million records”. It also advertises a catalogue of “older, rare and specialist volumes”.
It also comes at a time when debates around copyright and fair use are at an all time high – just last week, Anthropic was ordered to pay a landmark US$1.5 billion settlement to authors whose work was unlawfully obtained for the same purpose.
Stay inspired, follow us.
