Is There a Danger of Jumping the Gun?

Do you think there's sort of a danger of like false starts and whether it's at the public policy level or even in the private sector area? It's possible that AI can be a jerk, right? I mean, we've definitely seen examples of them being jerks. The real danger is that generally they're idiots, right? And they're making important decisions. That's why human in the loop is really important to me, especially in cases where it's high risk.