3:00 PM via Zoom
ABSTRACT: “AI Can Now Detect Depression From Your Voice,” reads the headline of a 2021 Forbes article, “And It’s Twice As Accurate As Human Practitioners.” This statement emblematizes a growing, primarily US-based subfield called vocal biomarker research, in which engineers, computer scientists, and mental health care professionals collaborate to develop technologies that they hope can identify biological indicators of mental distress expressed involuntarily in the sounds of the voice. The various stakeholders invested in vocal biomarker research tend to classify their technologies as “clinical-decision support tools.” Their aim is to augment administrative, para- and pre-clinical practices of discernment such as screening and triaging, rather than diagnosis or treatment, which they frame as indelibly human interactional arenas that cannot be replicated by machines. Meanwhile, they position screening as American mental health care’s most vexing and viably automatable problem-space because it is too human: too subjective, non-standardized, and sociocultural. This talk unspools the racial and gendered imaginaries animating vocal biomarker research’s multiple “genres” (Wynter 2003) of human and human-like listening subjects. I draw from ethnographic fieldwork with vocal biomarker research labs and human research subjects. Focusing on one lab’s efforts to craft the anthropomorphic user interface of a vocal biomarker technology, I argue that one of the subfield’s social effects is the enactment of a racialized and gendered hierarchy of listening. Nevertheless, research subjects’ subversive interactions with the interface highlight the instability of this hierarchy, while also destabilizing another key premise of vocal biomarker research: the universalizable mentally ill speaking subject.
BIOGRAPHY: Beth Semel is an Assistant Professor of Anthropology at Princeton University and an ethnographer of science, technology, and language. She studies the sociopolitical life of machine listening technologies in US mental health care research and practice, drawing from feminist science and technology studies, linguistic and medical anthropology, critical algorithm studies, and disability studies. Prior to arriving at Princeton, she co-founded and served as the associate director of the Language and Technology Lab at the Massachusetts Institute of Technology, where she received her PhD in History, Anthropology, Science, Technology and Society in 2019.