Protection from algorithmic bias? Really? It's not ideal that algorithmic biases exist - and to the extent that it's feasible unbiased algorithms should be use

Protection from algorithmic bias? Really? It's not ideal that algorithmic biases exist - and to the extent that it's feasible unbiased algorithms should be used, and mitigations provided for biases that cannot be avoided completely. However, it's not always possible - and we probably shouldn't avoid all progress just because it cannot be universally applied. In image recognition (within the visible spectrum) people of dark complexion will always have an algorithmic bias against them - as will any relatively low contrast picture. This isn't the result of racial biases in researchers, or the data sets, or the ever oppressive totalitarian system - black people reflect less light so their features are harder to distinguish. There's simply no possibility for any racially unbiased facial recognition algorithms. I don't know the solution here, but it should be a considered response and not a knee jerk reaction.