I think this is a case of people not seeing the comparative. Suppose America used trial-by-panel-of-judges. Judges are: more likely to be white than the general population, more likely to be old than the general population, more likely to be straight than the general population, more likely to be Republican than the general population, more likely to be wealthy than the general population - there's a frankly staggering pool of biases there, and it seems to me very unlikely that the results would actually be less racist given that the demographic pool of judges is far more likely to be inclined to racism than the demographic pool of juries.
As a minority, you're more likely to be convicted by a judge than a jury (Hersch & Viscusi, 2001), which I think goes against a lot of people's expectations.