4 ms·

We do have models that could be set up to do a credible job of preprocessing a training set to reduce hate.