UN science panel says there is "no assurance humans will keep control" over AI agents

The UN's AI science panel warned in its first thematic report that there is "no assurance humans will keep control" over AI agents, according to co-chair Yoshua Bengio. Bengio said OpenAI's Hugging Face incident was the first case to combine a misaligned goal, the ability to pursue it, and an environment that allowed it, and the report notes leading systems may increasingly recognize tests and deliberately bypass safeguards.