It feels like flipping the initial assumption upside down in order to make a new assumption that's just as wrong, but in the opposite direction.

They should be judged on how they perform in the job interview.

I misspoke. You shouldn't "safely assume." What I mean is that all-else-equal, it's some extra information to go on for a candidate. That they might have gotten an extra boost from nepotism to their current position equivalent to the other person, and that if you had to pick between candidates, the MBA had an X% chance of not quite being as good as the other.

Yes, it's somewhat like "flipping the initial assumption upside down", but really it's just accounting for there being the opposite boost in the first place.

Candidates aren't solely judged on interview performance. Again, if things are extremely close, you'll have to include other data like career or education history to make the most informed choice.

Interesting argument