Ok so “misaligned AI” is dangerous. But if it does decide to kill us, what SPECIFICALLY could it do?
Loading summary