Ways of Machine Seeing as a Problem of Invisual Literacy

The essay takes its point of departure in John Berger’s Ways of Seeing (1972) to examine how machine vision further unsettles what we see and what we know. When the majority of images are made by machines for other machines, and distributed across planetary networks, and part of vast annotated datasets, how are worldviews reinforced differently, and what kind of literacy applies, if at all? To clarify what is meant by literacy, the essay refers to its roots in cultural studies, to expanded notions of such as coding literacy. But what about computational vision, and its particular way of seeing the world, and how to gain access to its underlying structures and effects? Looking at source code to understand how a machine sees is not particularly revealing in itself, but rather requires a more relational approach. Furthermore, it is not simply a case of how humans see the world anymore, or how they use machines to see (as in the case of photography), but how machines see and produce the world in their own terms. The essay argues that we need literacy now more than ever to understand how forms of privilege are reproduced and naturalised through new ways of seeing.