Frankenstein in the Pipeline: Computational Epistemicide in Facial Recognition
A new paper proposes "computational epistemicide" as a framework to understand how facial recognition systems destroy the face as a living, relational surface. The research argues that by progressively narrowing the face to data points through processes like detection, landmarking, and embedding, these systems create a numerical proxy for identity. This vectorization process reassembles the dissected face into an artifact that circulates across databases, making recognition inseparable from standardization and rendering reformist "ethical AI" insufficient. AI
IMPACT Proposes a critical framework for understanding the harms of facial recognition, suggesting abolition as a normative stance against vectorized identity.