There is an argument to be made that you shouldn't be treated differently based on the color of your skin. In fact it is e.g. in the German constitution.

http://m.spiegel.de/international/germany/racial-profiling-g...

For example, if blue people are twice as likely to do bad things as red people, you end up catching more than twice as many blue people when you profile based on blueness. And that's unfair discrimination. It's unfair on all the honest blue people made to feel like criminals.

And of course it's self-perpetuating. The lopsided catch rates justify profiling in the minds of people doing the catching.

In the real world, I dont know the underlying actual differences in offending rates, but I do know that correlation with life circumstances would probably be much higher than skin colour.

You got anything to back up this blanket statement?

It’s just a type of argument that is actually quite common with its basic idea being that just because something is effective, doesn’t mean it’s right or just to do.

Transferred to racial profiling the argument is this: Even if racial profiling is effective, employing it is still ethically indefensible and plain wrong.

Put another way: Even if we agree on a goal (e.g. catching more criminals) that doesn’t mean that any effective way to achieve or improve that is then suddenly appropriate.

The German constitutional court does make this argument explicitly (even though I’m not sure any police officers actually respect that – but at least those affected can now somewhat defend themselves).

With arguments, and to support your point of view.

Maybe I wasn't clear. I wanted some real evidence that racial profiling is effective (as well as how effective it is), I wasn't looking to get into an argument about the morality of the issue.

Especially in the case I cited, it is obvious that racial profiling works better than random sampling (or based only on other features). People out of an ethnic minority will always have different rates of citizenships compared to the general population. So if you pick someone from a minority group that are less likely to be citizens, your chances are higher to pick someone who doesn't have the citizenship compared to the general population of citizens.

While the principle is quite simple, I don't have any empirical knowledge and can only refer you to Google Scholar if you desire more insights ;)

Let's say that Arab men between the age of 20 and 35 were most likely to steal something when they entered the store. (Not anecdotally: let's say the location computed the demographics of thieves going back 20 years, and found a consistent trend. Let's assume it's an objective fact.) Would it be okay to target them because they proved to be the highest risk?

I think the answer is no, and I think the answer is no also in the case of white teenagers. Sometimes doing what is most effective or efficient isn't the right thing or the fair thing.

Which is precisely why blacks are "profiled" or "targetted" even by computers. And yet the media tells us this is because "AI is racist" and because of the "legacy of colonialism" and absurd nonsense like that.

The issue is that the pattern matching reinforces the problem based on race when race is not the cause. There are lots of historical reasons someone could be be a bad credit risk for example, but it isn’t because they’re black or some other race.

You want the algorithm to account for cause not correlated patterns that reinforce racial divisions and then perpetuate the inequality.

It’s not because of the color of someone’s skin. It’s not even clear where racial lines are drawn and previously distinct “races” have disappeared into the general “Caucasian” label depending on time period.

Your comments are why it’s a problem - people are unable to understand the difference and overfit to patterns in the data. Our learning shares a similar problem to ML.

When you don’t correct for this it hurts people of that race and helps to perpetuate the imbalance - reinenforcing the historical data and increasing inequality.

https://www.cell.com/ajhg/pdf/S0002-9297%2818%2930363-X.pdf

"Genetics demonstrates that humans cannot be divided into biologically distinct subcategories."

"Most human genetic variation is distributed as a gradient, so distinct boundaries between population groups cannot be accurately assigned."

etc.etc.

You can only make that claim by examining a single locus. When you examine multiple loci, human populations are clearly separated into distinct groups. The fact that this oft repeated myth is so commonly addressed that it is a named fallacy should be an indication that it warrants a little investigation into its accuracy. Here's a handy graph to help visualize how Lewontin's claim is dishonest and false: https://ars.els-cdn.com/content/image/1-s2.0-S03069877090053...

If you want to argue about the science, please research the subject first. There's no way anyone could do even a cursory bit of research and not come across this as it has been beaten to death, resurrected, and re-beaten to death.