Contribution short abstract:

An exploratory anthropological investigation of the biopolitics of AI-powered emotion recognition systems in behavioural training programs in childhood education and for individuals on the Autism spectrum.

Contribution long abstract:

Anthropologists have long been attuned to the epistemological gap between how bodies feel and how individuals make sense of how they feel (White 2019). Diversifying our understandings of the spectrum of human emotion and psychic life is part of our discipline’s mandate. Psychological anthropology is urgently needed to critically evaluate the design and development of emotional recognition technologies. In the field of affective computing, computer scientists are designing emotion recognition technologies to facilitate the emotional and social development of grade school students and individuals on the autism spectrum (Picard 2020; Aylett 2018). To be digitally scalable, current innovations in affective computing rely on a highly reductive model of emotion theory grounded in Western psychological science (Barrett 2019). Affective computing labs have become speculative experimental sites that attempt to compute and mimic human feeling. Yet, the technologies they deploy run the risk of reducing any “emotional” divergence into Western norms of standardized behaviour and increasing the social stigmatization of Autism Spectrum Disorder. This paper seeks to examine the biopsychic imaginaries and biopolitics (Stevenson 2014) surrounding the engineering of AI-powered emotional recognition technologies and discuss the consequences of attributing caring and teaching roles to digital companions.