Can AI agents be made safe as they become more autonomous?

AI agents act autonomously to complete goals, but experts warn their “black box” reasoning makes safety hard to guarantee.