Earlier quoted context omitted.
> The problem with erasing biases is that you cannot look at any statistics. The internet and training set can be free of any form of -ism, and the models would still be expected to have biases. In fact it's something desirable, because statistical inferences are a valuable tool. I’m sorry but I can’t let you get away with this terrible argument and conclusion. No one argues for completely erasing bias (especially th…
Ok, but what if you're a non-emotional system, whose biases are generated from objective statistical data? Then you become aware of those biases, and "adjust our opinions and behavior". You are just introducing inefficiency, and it speaks to the OPs point of: > The problem with erasing biases is that you cannot look at any statistics. If you can't use the statistics to generate biases then what is the purpose of buil…
Objective statistical data doesn’t exist, that’s Data Science / Statistics 101. Your sample always has a bias, unless your sample is: everything, always, how it’s been, and how it always will be.
I don’t really know what inefficiency has to do with anything, wish I could respond to the rest of your comment.