Overall these seem like very reasonable safeguards. What I'd like to see very clear regulation around is applications of AI in particular domains. I don't care

Overall these seem like very reasonable safeguards. What I'd like to see very clear regulation around is applications of AI in particular domains. I don't care that an LLM failed to analyze a document correctly on my behalf. I care quite deeply about using it to make a decision in a high-risk and/or high consequence (to human life) environment. I care a lot about having the bias inherent in any model drive a decision without human intervention.