The AI lab waging a guerrilla warfare over exploitative AI

But it’s “simplistic to assume that when you’ve got an actual safety drawback within the wild…