Same with machine learning and statistics.

There's been a couple of notable cases where "using AI to solve X" turns out to be "X immediately arrives at the correct (but highly politically incorrect) conclusions that people already know... now panic and backtrack because political correctness".

I do wonder how much of the current flailing of some companies is due to some combination of avoiding statisticians in general (due to the potential of arriving at correct (but highly politically incorrect) conclusions), and avoiding good statisticians in particular (due to their annoying habit of sticking by correct conclusions even when they are politically inconvenient).