Agreed. We can't expect human behavior to change, because it won't. We need to design safer systems instead.

The only "law" I agree with is:

> Humans must remain fully responsible and accountable for consequences arising from the use of AI systems.

And that starts with framing, especially in the clickbait "AI deleted the prod database" headlines. Maybe we just start with saying "careless developer deleted prod" because really, they did. Careless use of a tool is firmly the fault of the human.