And I must write my truth before I get to anyone else’s
Racially biased algorithms operate according to a predetermined set of racist statements given by the creator of the system. Biological correction factors for socially defined categories such as race lead us to the place where we take race as fact. A social definition that is not universally upheld can never be directly translated into a number unless it is bringing aside a qualitative analysis.
Example: “Black people are more likely to be exposed to gun violence.”
Bad Example: “Black people are genetically pre-disposed to violence.
While the first sets up the rest of the essay as an observed scenario and spend time questioning how and why the phenomenon occurs, the second leaves people to know the conclusion of the statement.
But how do we get to this point?
We have social problems attacked in areas that are not ethnically and ethically conscious. Take this for example. The first time I learned about Machine Learning, it was from a man who said that anti-Black racism is not real. He argued that the problems of the Black community were rooted in the lack of father and parenting issues. He then went on to say that most Black people didn’t have dads without citing it. Despite showing him numerical evidence of how oppression is an upstream process and acknowledging the issues of Black families can’t be brought up without mentioning high arrest rates, social welfare positions, and cultural assimilation, he didn’t believe me.
He didn’t believe the simple fact that racial disparities came from years of oppression extending from slavery. He insisted the Black behavior was the root of Black plight.
Citing me as an example of rags to riches. Coming from a home where I spent years with my birth father in and out of jail and my step father separated in my mother, my identity guilt rendered me weak. It was hard to argue with someone so blatant and reckless about your life.
He was confident, intellectual and had a Bachelor’s degree in Bioengineering. He was a master’s student in Business in worked in a Brain fiber tractography lab, focusing on the software for producing 3D images of brain scans. He had access to patient profiles, including demographic data. He was also a minority, multiracial, and an immigrant.
The people producing high volumes of work that rely on racist assumptions are all academics. They are respected and often not questioned, as bioengineering is a field that hides from race.
A corpus callosum autosegmented through a machine learning fiber tractography project.
This is someone with the potential to misuse their multiracial identity as a reason to teach that Black people are naturally inferior. That is exactly what he did to me, and I decided to stomach it so that I could stay open to learning. He taught me about classification issues that would inhibit training in supervised ML. He taught me how easy it was to be reckless with language when no one was watching. Of course I was the only Black student in class. Of course we had these conversations when no one else was around. This man is one career breakthrough away from pathologizing Blackness without acknowledging racism.
Do I believe he meant harm? Absolutely not. But the consequences of his behavior are disastrous.