Tracing the Contours of Algorithmic Care: Logics of “Real-Time” Mental Health Interventions

• In Event: Interrogating Innovation in Global Digital Mental Health

Abstract

Recently, with the application of artificial intelligence (AI) and machine learning (ML) to a wider range of healthcare contexts, there has been a rise in computational systems that aim to provide real-time mental health support and advice. Facebook’s suicide prediction tools, for instance, analyze the site’s social media posts seeking particular patterns across text and video data with the goal of stopping people who are at risk of taking their life by triggering a range of online and in-person interventions. While the privacy and ethical implications of such technologies are currently being debated, this paper draws attention to the cultural and computational logics that underlie real-time mental health interventions by examining the type of algorithmic care work enabled by the design of systems centered around ‘‘safety,’ ‘speed,’ and ‘scale.’ Care, as techno-feminist scholars like Annemarie Mol and María Puig de la Bellacasa have shown, are bound up in particular epistemologies, sociomaterial relations, and biopolitical configurations that need to be critically examined. Placing techno-utopian narratives of “AI for good” in conversation with the techniques of AI/ML and ethnographic research on the lived experience of people living with chronic behavioral health conditions in the American Midwest, this paper seeks to outline the sociotechnical boundaries of “real-time” or “just-in-time” digital interventions and who is left out of such a vision of care.

Author

• Elizabeth Kaziunas, AI Now Institute, NYU