When Technology says, "I Don't See Color"

We've all heard it before. I've heard that phrase, "I don't see color", one too many times. That phrase elicits a reaction that seems to make my entire body cringe. Many people, especially those of color like me, hear that and think, "Seriously?" or they may react like, "Sure....ok.". I mean, how could I possibly believe that? As a black man, I have been forced through my experiences in life to understand how the structures of society have been stacked against me. There was never any choice. I had to witness first hand how schools, businesses, and other institutions have not been built for my success, but for me to fail instead with little to no helping hands all due to the color of my skin.  In my life, I've witnessed how racial biases have led to instances of racist profiling, discrimination, and in extreme cases, violence or death against people of color. So now, what happens when you have technology that is discriminating against us? How is it possible that our tools even acquire bias?

image from article: Facial recognition isn’t always accurate, nor unbiased.

In Ihudiya Finda Ogbonnaya-Ogburu's case study of "Critical Race Theory for HCI", she explains how biased technology focuses on "algorithmic bias – systemic errors in computer systems that lead to unfair outcomes or judgments". Some systems struggle or don't recognize darker skin tones. There have been instances of commercial facial recognition technology that was found to be more accurate for white faces than for black faces. These types of biases have also led to real-world consequences, for example, in terms of how law enforcement have used facial recognition. Another example can be mobile apps like Pokemon Go for instance. One would think that these type of cases would be harmless. However, for apps like Pokémon Go, or apps with delivery algorithms, the areas of availability are another example of racism in technology. Areas such as minority neighborhoods suffer from redlining. This means that they’re excluded from participation due to their location, which is why there are fewer PokéStops in minority neighborhoods.

Image from USA Today

These types of critical errors should not be acceptable. Yet, many other examples of technologies exhibiting racial bias have come up numerous times. 

Now, solely blaming the technology for these flaws is inexcusable. AI is not going to pick up on the fact that it is being fed bias or alert us of biased results because it doesn't have a moral compass. It doesn't know any better. We also must remember that these are technologies that were built by us. Philip Brey wrote that to better keep these technologies in check, we must, "morally evaluate practices that involve ICT and to devise ethical policies for these practices." Technologies like AI and machine learning have been proven to show error in regards to racial biases. We must have the wherewithal to hold the morality of our technologies accountable along with the individuals who help to build them. We can't afford to continue to be naive to these issues or we will continue to see more instances of technology causing unwarranted harm to people of color. Not acknowledging the risk we have at hand will only cause us to suffer the consequences, great and small.