The Importance of Balance in AI

I don't think that the call for a pause in AI research is a very good idea if only for practical reasons. I'm much more interested in figuring out what are the pathways for AI to actually do bad things and then try to figure out how to prevent those pathways from coming true. Craig Stevens says, my understanding is that one of the first demonstrations of the superiority of Einstein's theory of general relativity was that it correctly predicted the orbit of Mercury, which Newton's theory did not.