[SpicyLemonZest]

One example is unconscious bias training. Modern meta-analyses strongly suggest that it doesn't work at all, that whatever the IAT is measuring is unrelated to patterns of discriminatory behavior. This is well-documented in the literature, and by a subset of news sites including some left leaning ones. But it would be completely impossible for a big company's HR department to say "we are cancelling our unconscious bias training because it doesn't do anything".