AI safety warnings mount as frontier models test new limits - KATU

Experts are warning that the newest, most powerful AI systems are becoming harder to control and may pose safety risks that aren't fully understood. These concerns matter to everyone as these systems become more widely used in society.