Affiliation at time of award:
Department of Anthropology
Massachusetts Institute of Technology
Beth Semel by Elena Sobrino
Speech, Signal, Symptom: Remaking Psychiatric Diagnosis in the Age of Artificial Listening
Ms. Semel’s project ethnographically investigates the work of U.S. university-based research teams of psychiatric and engineering professionals collaborating to develop artificial intelligence (AI)-enabled speech analysis technologies. Researchers hope these technologies can resolve psychiatry’s longest standing issue: the subjective nature of diagnosis. Emphasizing the way culture shapes statements about AI’s “objective” capacities, Semel explores how researchers insist that their technologies can identify signs of mental illness that are otherwise inaudible to humans, and will be agnostic to some differences – like race, gender, and class – but not to the difference between a psychologically well and unwell person. What assumptions about mental illness and the relationship between language, listening, mind, and self are solidified in these diagnostic technologies? What do they reveal about the value of care in mental health care practices as partnerships between psychiatry and engineering become increasingly common in the United States? Semel argues that researchers’ attempts to use AI to cut through the sociocultural aspects of language in order to capture supposedly universal, biological aspects of mental illness reflect ambivalent attitudes about how best to listen to patients’ speech, and are part of a broader reconfiguration of expertise within U.S. mental health professions
Generous funding for this Fellowship provided by the Weatherhead Endowment.