"CriminalBlackMan In Code" 2026
Projection onto warehouse wall
Runtime: 1:42
Film comprised of edited booking photographs of black men that were considered to be visual "matches" to an unrelated image of my long lost high school friend. We'll call him "V"
Projection onto warehouse wall
Runtime: 1:42
Film comprised of edited booking photographs of black men that were considered to be visual "matches" to an unrelated image of my long lost high school friend. We'll call him "V"
I had searched for years to find "V". After almost a decade, I found an image of his mug shot taken around 9 years after we had graduated.
After doing a reverse image search of his face, Google results showed me several other black men who were also arrested on miscellaneous charges. They looked absolutely nothing like "V" at all.
Google results claim these are the same men
The name of the work is in reference to the book "The Color of Crime: Racial Hoaxes, White Fear, Black Protectionism, Police Harassment, and Other Macroaggressions" by Katheryn Russell-Brown originally published in 1998. This is where she originally coined the term "Criminalblackman" as a sociological phenomenon that lives on in the American psyche through the continuation of racial biases in media and culture.
The technology in facial recognition systems often use "Booking" data and not "convicted" data. This allows for the visual information subset to be inherently biased towards black men who are already disproportionately arrested under false pretenses. That Data is then used to train AI Algorithms to match suspects to civilians. I have found two separate cases in the last 5 years that have falsely matched two Detroiters with crimes they had not committed.
See Woodruff v.Oliver and Williams v. Detroit
In addition to poorly conducted data collecting practices, AI Technology struggles to recognize black facial features. In 2019, a NIST study discovered that Black and Asian Faces were up to 100 times more likely to be misidentified in comparison to white faces.
See Woodruff v.Oliver and Williams v. Detroit
In addition to poorly conducted data collecting practices, AI Technology struggles to recognize black facial features. In 2019, a NIST study discovered that Black and Asian Faces were up to 100 times more likely to be misidentified in comparison to white faces.
Nearly 30 years after the first edition of "The Color of Crime" media technology and surveillance technology work hand in hand, continuing to push false narratives and governing with poorly organized data without legislation or guardrails.
These results are inevitable. What exists is by design.
These results are inevitable. What exists is by design.
Sources:
ACLU
The Guardian: Facial Recognition technology discriminates against people of Color
University of Calgary
BBC
Harvard
Cases of Facial Recognition software misidentifying black men:
PBS
UMICH
ACLU
The Guardian: Facial Recognition technology discriminates against people of Color
University of Calgary
BBC
Harvard
Cases of Facial Recognition software misidentifying black men:
PBS
UMICH
