Data-based AI, rather than being neutral, seems to be a tool for amplifying and extending human bias. Whereas a person might occasionally treat a member of a di

Data-based AI, rather than being neutral, seems to be a tool for amplifying and extending human bias. Whereas a person might occasionally treat a member of a discriminated against class neutrally, whether by accident, oversight, training, human decency, moral bias or a different form of discrimination - on the other hand, a computer will always do the same thing, and amplify the "people from poor backgrounds are refused entry to places that might increase their wealth". Police get more arrests where there's more police, so if you use an AI to work out where to send cops, it's going to find you the place your discrimination wants it to. The modern data-ocracy is just the old eliteocracy with a new coat of paint.