Top official warns against unchecked global AI competition, corporate safety commitments, AI safety and national security, frontier risks in technical standards plans, papers on superalignment