I think the strong counter-argument to this is in [1]. Quite simply we need to make sure we don't get fooled by randomness and see bias where there is none, thereby making good algorithms worse in an attempt to fix imaginary bias.
I don't think it's that good counterargument. The explained problem was treated as simple mapping - this value causes that result. It spends a bit talking about what statistics mean, but unfortunately doesn't discuss the idea of confounding variables, which influence both sides of the equation.
[1]: https://jacobitemag.com/2017/08/29/a-i-bias-doesnt-mean-what...