In Canada, we are only 3.5% of the population but are somehow stopped at twice the rate of white citizens. I would love to slay your argument with numbers but since Canada refuses to collect data by demographic (which I suspect is done to keep the misinformation hidden) I can't provide solid numbers.

In Europe, they combine Black, Asian and all other minority ethnic groups , which of course is to skew the data. A simple breakdown of that combined group will show you YET AGAIN that you are absolutely incorrect. https://irr.org.uk/research/statistics/criminal-justice/

>Black people do not commit way more crime. How is that even statistically possible when we are only 12% of the population

I mean, the AI would look at the relative rate of crime between population groups as part of the process of discriminating who should get a resource to maximize whatever outcome its looking for. Black people are frequently low income and urban, and are part of gangs at a higher rate for that reason. Besides the whole policing bias issue, where people are concerned about black people being over represented in stats because the police are patroling around them more often, for racist reasons, there are a lot of signs that gang violence is overrepresented and indicates an actual objective inclination towards a much higher rate of crime that will for sure come out in the data.

Murders for example are hard to cover up because you have an actual body on the ground, and the black crime rate for murder is 50X any other population, so an AI will look at that and draw conclusions. It seems unlikely that the white or asian population have 50X murders going on in secret that they just cover up really well using racist police. Doesn't mean that black people all murder, but any policy from the AI is necessarily going to discriminate against the group as a whole, unless it were to have perfect information about every individual.

I mean, that table you reference in the link says White 45% murders, black 51% murders.