An AI safety expert outlined a range of speculative doomsday scenarios, from weaponization to power-seeking behavior

A recent research paper considers the potential future dangers of unfettered research into AI — and the scenarios are nightmarish.