When a measure becomes a target, it ceases to be a good measure.
Think carefully about if racial equality is truly your end goal. I agree that in a perfect world, there would be racial equality. I also agree that there were, and are, forces pushing us away from racial equality and that those forces are harmful.
Consider two children, one black and one white, who are in an approximately equal position. Why is it beneficial to give advantages to the black child just because his condition was caused by generations of discrimination, instead of, say, crappy parents. Perhaps it would be easier to consider the population of "low class" people [0]. Under what theory is it beneficial to discriminate in within this population?
This is not to say that, from a practical perspective, affirmative action is always wrong. For instance, if you have a white applicant and a black applicant that, by all available metrics, are equal; it is entirely possible that the black applicant is actually a better choice for reason that statistically exist, but which you fail to capture in your metrics. But, unless you are arguing that there are actually meaningful differences between the races, you are simply using race as a proxy for some measurable property(s). It would be better if we could measure that property directly (although, we often cannot).
If we simply force racial equality without address the root causes of the inequality, then we have two problems: The first problem is the harm caused by the initial racism. The second problem is the harm caused by the racism inherent in our method of forcing equality. There is also the benefit caused by incidentally addressing the root cause of the inequality; but since that is not what we are measuring, it is not clear how well we are doing with that.
[0] Yes, in general low class whites are better off than low class blacks. But the curves have enough overlap that it is possible to construct a population where the expected value of a white person is equal to that of a black person. That is the type of population I am talking about.