What are the best ideas of how to regulate AI from the US executive branch?

I'm of course interested in trying to prevent catastrophic accidents from AGI. However, likely the more immediate and more tractable potential regulations are much nearer-term and clearly aligned with current US government incentives, eg. governing the development of autonomous weapons systems.

Also, I'm interested not only in the ideas, but also in the implementation of those ideas. What departments and teams are working on these? In order to make these ideas reality, what needs to happen - who needs to take what actions?

Any thoughts you have would be useful, as it will help me with the capstone project for the AGI Safety Fundamentals course (more info here). Thanks in advance!