Assuming performance on the SAT is an unbiased measure of ability is the flaw in this premise. It is well known that SAT performance is correlated with race, fa

Assuming performance on the SAT is an unbiased measure of ability is the flaw in this premise. It is well known that SAT performance is correlated with race, family income and gender. For instance, girls reliably score worse on SAT Math even when they outperform boys in school on the same subject. However, it turns out family income is an even stronger predictor of test performance. But family income also happens to be correlated with race, so the race/SAT score correlation is more likely an income/SAT score correlation. In fact, some even maintain that the process of selecting SAT questions is itself a self-reinforcing bias. If you want data, searches like "Is the SAT biased" or "Is the SAT racist" will take you down that rabbithole. Given that the testing process has clear shortcomings, it seems fair to account for that during admissions. Unfortunately, this looks a lot like very obvious reverse discrimination if one is not aware of the non-obvious, systemic discrimination it is adjusting for.