AA can totally fuck up if it's applied as a band-aid at the final stage of the pipeline however.
Any sort of diversification effort needs to apply small biases at each stage in the pipeline, rather than one big bias right at the output end.
For example, say that you've decided that <target group> is seriously under-represented in your <education course>. So, you apply an aggressive Affirmative Action policy, by giving some sort of bonus points to the under-represented minority. So, therefore the number of <target group> is instantly increased to be representative. Quick fix solved all the problems, right?
Well not really. If there was any type of merit-based entrance criteria, you'd have to exclude e.g. the worst-scoring white people, and bring in a large number of black people who scored below the "previous worse" point. This means you've engineered a significant gap between the measured ability level of the lowest-remaining white person and the best-performing black person in the new admissions group. The net result is that the "average score" of white people in the class
increased (because you excluded the bottom-scoring whites), and the average score of black people
decreased (because you admitted a significant number of even lower-scoring blacks). This causes a racial division in average scores within the classroom, which can exacerbate or even create stereotypes about ability vs race, even when none existed before.
With TV writers, I'd ask, what's the racial breakdown of college-level writing courses? Does the hiring of professional writers mirror what's coming out of graduate schools?
https://www.clevelandfed.org/newsroom-and-events/publications/economic-trends/2015-economic-trends/et-20150331-racial-and-ethnic-differences-in-college-major-choice.aspxIf you take a look there, only about 2.8% of English majors are black in American universities. People can't hire people who aren't doing the appropriate courses of study.