bias training is a fun phrase and aptly criticized.... you can train ai (and presumably large primates) with it. It will yield to the bias induction, but more importantly than that, it will deduce it as being the only possible truth, based on a slimming margin of error on a probability check. It's fairly accurate, but nowhere near the sort of logic required for critical morality or survival decisions. As a practicing human, it seems rather obvious that it would be bad to expose us to 'bias training' of any sort.