It’s probably meaningful that I come across this article on Urban surveillance networks and this article on profiling on the same day. Both of them are worth reading.
Consider the following statement.
An overwhelming majority of Americans think that racial profiling is wrong. A lesser number think that racial profiling is not worth doing under ordinary circumstances, a smaller number think it’s not worth doing under any circumstance.
The above is an accurate description of public sentiment when 100% of the factor is race. Gender and age are usually thrown in as well. The above still holds true.
But what happens when race is one factor of 50 and the profiling is being done by a computer? Assume a surveillance server can determine, height, weight, approximate age, race, gender, posture, gait, clothes, et al. Does it become acceptable at that point?
This line of thought reminds of the last Supreme Court affirmative action decision where it was said that it was wrong for people to discriminate, but fine for computers to do so.