Comment by ETH_start

That is simply not true. We have accountability methods employed that are themselves AI-assisted, that help us gauge the alignment of various AIs.