I did mention that it was written by a human, yes, but the reintegration part is called “machine learning” and doesn’t necessarily require any further human input once the algorithm is given its authority.

I’m trying to say the racist outcome in this example isn’t the result of some tyranny of numbers that we need to keep subjugated to human sentiment or something. It’s actually the result of human overconfidence in the future mandate of our technological achievements, which is an emotional flaw, rather than something inherent to their simple performance as tools.