Algorithmic Images and Recursive Epidermalization

In this chapter, I apply this concept of the photographic image’s dissolution into social architecture to the advent of photorealistic images produced using generative artificial intelligence and machine learning techniques. Methodologically, I connect notions of imagistic social architecture derived from Benjamin and Vilém Flusser to Frantz Fanon’s discourses on sociogeny and epidermalization. The former refers to the ways linguistic and cultural norms come to be experienced as “instinct” to such a degree that they can rewire the mappings between sense impressions and biochemical reward/punishment pathways in the human body, the latter to how subaltern, dark-skinned, postcolonial subjects can internalize normative cultural assertions of their inferiority. This is an urgent juxtaposition because the memetic, appropriative qualities of algorithmic images bind them more closely to hegemonic ideologies and normative skin color stereotypes than film photography, which maintained a relationship to real events due to its semi-indexical relationship to light rays. In other words, I argue that algorithmic images produced using generative techniques are necessarily more “biased” than the photographs they imitate. Further, they reify race and color stereotypes as a direct result of their grounding in counterfeiting processes, where “believability” replaces “indexicality” as social authentication mechanism.