ProPublica documented that among defendants who did not reoffend, Black defendants were flagged as high-risk at a higher rate than white defendants. This is a false positive rate disparity, not overall prediction rate (A), calibration failure (C), individual fairness (D), or treatment equality (E).