'Dystopian' - AI firms target London’s rare bookshops in hunt for training data

AI's hunger for data is now consuming rare books and booksellers are cashing in

AI firms bulk-purchase rare books (millions per order), scan for LLM training, destroy originals—targeting pre-2022 clean data. Exposes training-data quality as a foundation-model differentiator, validating premium-dataset pricing and reshaping LLM procurement—while escalating IP/regulatory risks.