Is AI really getting dangerously out of control?

Grim warnings of the existential risk posed by AI are all over the news, but there is very little evidence to back them up