I think it's more important to think about how we can support responsible AI than to focus too much on the penalties, because there are socially beneficial applications of AI. Openness in science has led to a lot of improvements in AI and how responsible it is.
I would focus on that aspect in particular.