Not sure people trust or desire formal reasoning enough for this to be likely. I’d bet we make due with informal reasoning and rationalizing, dividual AI —i.e. as with humans, giving plausible/simple/ethical reasons is necessary, but may have nothing to with the initial decision.https://twitter.com/plinz/status/1006912797932969985 …
I don't think that "people" in this sense matter. Corporations, nation states and large organizations cannot afford to be guided by irrational impulses. Especially when they compete with rational agents.