Computer vision technologies routinely misclassify or fail to recognize nonwhite faces, and many have argued that this indicates a lack of diversity in the datasets used to train these technologies as well as a lack of diversity in computer programming teams. But Amaro offers a different kind of critique, one aimed at the very nature of computer vision technologies. Amaro argues that the inclusion of black faces in datasets, and thus the creation of systems that recognize black faces, is not the only response to this problem. In fact, he argues that it is a response that might make many problems worse.
Given that computer vision technologies are based on a logic of "coherence," such technologies constantly aim to make sense of that which they see as incoherent. If whiteness is the norm, blackness is seen as incoherent, as a problem to be solved. But for Amaro, blackness is coherent in that it is "continually taking shape." If an algorithmic system perceives this as incoherence and thus tries to either make it cohere (by comparing it against a norm) or excludes it altogether (further solidifying the norm), then the answer is not to incorporate blackness into that system. Instead, it is to imagine an entirely new system.
In this short writing assignment, your task is to imagine a technology inspired by Amaro's framework. What technology can you imagine that is not rooted in "coherence and detectability"? This does not have to be a computer vision technology, though it can be. Your job is to invent, to treat this as a thought experiment: What other possibilities can we imagine if we take Amaro's argument as our starting point? What other futures are possible if the "black technical object" is not understood as needing to conform to the algorithm but is instead understood as an invitation to think differently about technology and design?
