Solving the Conflict Between Breathability and Masked Faces within Facial Recognition Technologies

Abstract

The interlocking crises of the COVID-19 pandemic, wildfires exacerbated by climate catastrophe, and systemic racism in law enforcement have led to the biopolitical management of breathable air alongside the exponential rise in the production of masked faces. However, the introduction of masked faces momentarily short-circuited the abilities of facial recognition technologies (FRTs) to identify and verify the novel combination of media, which includes the elemental media of air and breath, that constructs masked faces. FRTs’ initial failures related to masked faces speaks to the internal conflict within contemporary biopolitical control that arises during moments of crisis, wherein previous protocols and practices clash with new material and media. In response to the competing vectors brought about by the crisis of breathability, the National Institute of Standards and Technology’s July 2020 report details FRTs’ failures as they relate to masked faces, while providing protocols for future solutions. This paper critiques the solution provided within the report in which the creation and deployment of synthetic facial data relies on the liminal populations of refugees and immigrants for experimental materials, a practice that ultimately targets the very populations being used to improve the technology.