Artificial intelligence companies are increasingly purchasing old and\nrare books in bulk, only to destroy them after scanning their contents\nto train AI systems. For developers of large language models (LLMs),\nthese physical books are viewed merely as raw material, with their\ncontent extracted and the original copies subsequently discarded. This\nmarks a shift from previous practices where companies often relied on\ndigital books or pirated copies.<\/p>\n
In a settled lawsuit involving the prominent AI firm Anthropic, it\nwas revealed that the company used a guillotine paper cutter to slice\nthe bindings off purchased books so the pages could be fed through\nhigh-speed scanners. The text is then used to train their AI systems.\nThis practice exploits legal provisions that allow purchasers to freely\nuse goods they have bought, treating the destructive scanning as a\ntransformative use that does not infringe copyright.<\/p>\n
Services like ISBNdb, which claims to have the world\u2019s largest book\ndatabase, are now facilitating these bulk purchases. The company argues\nthat print books published before 2022 contain cleaner, more reliable\ndata than internet text, which is increasingly contaminated by\nAI-generated content. It guarantees anonymity to its buyers to shield\nthem from public criticism.<\/p>\n
One book dealer noted a dramatic surge in sales, from dozens to\nhundreds of copies per week, with orders often identified only by book\nID numbers, strongly suggesting AI labs are the buyers. While the\nfinancial boon is welcome, the dealer expressed sadness that rare books,\nof which few copies remain, are being systematically destroyed and\nrendered unreadable for future human readers. The trend has ignited a\nfierce ethical debate over the destruction of cultural and scientific\nheritage to fuel technological advancement.<\/p>", "url": "https:\/\/jawawa.id\/newsitem\/ai-companies-destroy-rare-books-after-scanning-them-for-training-data-1785419546", "image": "" }, "sponsor": "Okusi Associates", "sponsor_url": "https:\/\/okusiassociates.com" }