It can be an interesting dilemma in practice. I supervised an honor group selection committee where the few male candidates who applied weren’t competitive enough in metrics to make the cut, potentially leading to a cohort of only women, our first.
The concern among faculty in the selection committee was that the freshmen community offered by the group would no longer be coed and the social dynamic would suffer from the lack of diversity. These concerns were raised primarily by women and NB faculty. The male faculty tended to prefer strict meritocracy. (I’m NB but didn’t have a vote unless there was a tie.)
Ultimately, the vote to alter our base metric weighting to allow a few men into the community failed. It did alter the community significantly, since the honor group gradually become a women’s club with occasional token men. Most of the honors students now seek out more diverse communities of other clubs by the end of their first semester.
I suspect this article makes two mistakes.
The first is treating GPA as the one true proxy of merit. The paradigm in admissions has shifted from simple GPA and standardized testing calculus to comparing like-with-like whenever possible. For example, the CV of an applicant who is a first generation college student is usually not easy to compare to that of an applicant whose parents are doctors, so admissions teams will often attempt to factor this into evaluation. The point isn’t discriminating against legacy students. It’s to recognize potential and ensure access and diversity of perspective.
The second is conflating different arguments for diversity. Affirmative action was established to help correct a long history of systemic inequality. And that’s why “affirmative action for men” sounds absurd. But inequality is not the only motivation for diversity.