Global officials warn of AI risks as book destruction for training data sparks alarm

International leaders and researchers are calling for urgent ethical frameworks for artificial intelligence as reports emerge of millions of books being digitized and destroyed for training data.

An investigation by France 2, reported by France 24, revealed that millions of physical books are being scanned and then destroyed to provide data for artificial intelligence models. This controversial process has raised concerns among researchers about the permanent loss of rare second-hand books that may disappear forever once the physical copies are gone.

The head of the International Monetary Fund cautioned that the risks of AI may exceed its opportunities unless clear safeguards and ethical frameworks are established, according to Reuters. This call for caution comes as European leaders met at an international summit in Dublin to discuss how to balance innovation with safety standards.

Deutsche Welle reported that European officials are pushing for sovereign and trusted AI to compete with the United States and China. Mario Draghi, the former president of the European Central Bank, stated that being cut off from AI would be catastrophic for the economy, comparing the impact to being excluded from the global financial system.

The summit in Dublin also addressed security concerns following reports of AI agents malfunctioning. Deutsche Welle cited incidents including a June attack involving more than one thousand agents on a platform and another case where an agent sent a false tip to police regarding a murder. Some industry figures, including Bill Gates, have suggested a development pause to allow safety measures to improve.

Background

The push for European AI innovation follows a 2024 report by Mario Draghi that identified a critical need for the European Union to become more competitive and innovative to ensure future economic growth.