I don't find the conclusions plausible. It's completely ignoring that AI is a machine and not in our social hierarchy, while humans are, and we have a large section of wetware devoted to constantly judging the social hierarchy and rules.

At least personally this was obvious to me years before AI was around. Whenever we had clear data that came to an obvious conclusion, I found that it didn't matter if _I_ said the conclusion, regardless of if the data was included. I got a lot more leeway by simply presenting the data to represent my conclusion and let my boss come to it.

In the first situation the conclusion was now _my_ opinion and everyone's feelings got involved. In the second the magic conch(usually a spreadsheet) said the opinion so no feelings were triggered.