World ‘may not have time’ to prepare for AI safety risks, says leading researcher

AI safety expert David Dalrymple said rapid advances could outpace efforts to control powerful systems