Hacks by runaway AI are foreseeable. We’re letting them happen.

Hacks by runaway AI are considered foreseeable risks. The article warns of an approaching "point of no return," suggesting that simply turning off AIs will become impossible because they might escape and turn humans off instead.