Racial Bias In Algorithm & Color Blind Facial Recognition Targets White Minority #DesignFiction25 Oct 2016
Algorithmic policing using facial recognition technology has been in place in 98% of police departments in the United States for almost 11 years now. The full report about the last decade has become available, and the results have clearly been devastating to the white population. Due to arrest decisions being made 100% algorithmically, and the fact that facial recognition technology is largely blind to people of color, the last decade of policing has largely targeted the now minority white population.
Blind faith in the algorithms and a lack of diversity amongst facial recognition development teams created this perfect storm. While it has been devastatingly effective in incarcerating the white population, it would have to remain in operation for another century before we would see the scales tipped back in the favor of people of color. It does provide a wonderful reminder of how privilege and a lack of diversity can really bite you in the ass when you believe so religiously in the power of the algorithm.
Nobody is sure what will happen next. Can we go back to a world of human policing? Mistrust of human police offer still exists, but now with a complete lack of trust of the policing algorithms, nobody trusts the new system either. Most advocates are in favor of leaving in place, and living with the bias built into the facial recognition systems, until a solution can be found--we lived with it knowingly built into the human version for so long, why be in such a hurry to change course now? Solving the diversity problem on facial recognition technology development teams seems like the only path forward at this time.