“The U.S. government is currently incapable of effectively addressing the safety risks associated with artificial intelligence.”