Mandatory reporting of AI risk is likely to be pretty useless, considering how poorly most people understand how LLMs work.. Will be interesting to see how they try to enforce it..