Comment by sokoloff

Want to bet that there was a round of human review applied after the algorithm ran where, at a minimum, employees judged to be high-performing (or otherwise favored) were saved even though they were initially flagged by the AI/data analysis?