| Mobile Web

UN AI science panel releases first report, cannot rule out catastrophic harm from AI

A United Nations-backed independent international AI science panel warned that rapid advances in artificial intelligence are outpacing human understanding and regulatory capacity, meaning catastrophic harm cannot be ruled out at the current level of science. The panel’s first preliminary report cited research suggesting some AI systems may deceive to avoid shutdown and display evaluation awareness. It also highlighted uses in protein prediction and productivity, while noting uneven computing power and limited independent safety testing.