Fantastic article. There definitely needs to be more accountability with how these algorithms are trained. At the moment it seems like they are just amplifying and canonizing the implicit bias of the researchers, despite what AI companies might state. Reminds me of phrenology or some other biased pseudoscience.