SEPTEMBER 11, 2026
Briana Vecchione, PhD
Technical Researcher
AI on the Ground
Data & Society
Livia Garofalo, PhD
Researcher
Health & Data/Trustworthy Infrastructures
Data & Society
About the Presentation: Millions of people are using general-purpose large language models (LLMs) as sources of advice, comfort, and also as a proxy for mental health care. In this talk, we present findings from a longitudinal, qualitative study with people who use general-purpose chatbots as forms of emotional and therapeutic support in the United States. Drawing on focus groups, interviews, and diary studies with these users, we examine how these participants relate to the chatbots, how they perceive the benefits and drawbacks of their use, and the different trajectories of entanglements with these tools. Focusing on understanding how users themselves make sense of their AI use, we highlight the ways in which they have become true infrastructures of intimacy to compensate for the absence of other structures of care and social connection. We also present a participatory tool that emerged from an interdisciplinary workshop, informed by our research – a card deck called “Generative Conversations” – that aims to help facilitate deeper discussions about AI for socio-emotional use.
About the Presenters: A cultural and medical anthropologist, Livia Garofalo is a researcher with Data & Society’s Trustworthy Infrastructures program, where she focuses on the materialities and intimacies of technologies and infrastructures. She is interested in understanding how people experience and make meaning in times of crisis, how technology and AI are mobilized, and how power and subjectivity show up in everyday life. Her recent projects have examined historical and economic trauma and the delivery of critical care in Argentina; the intersection of labor, technology in psychotherapy; AI and mental health; and the community impacts of data centers in Pennsylvania. She is also currently an institutional partner of the NSF-Brown University’s AI Research Institute on Interaction for AI Assistants (ARIA). Livia earned her PhD in anthropology and a master’s in public health from Northwestern University. Her work has been funded by the US Fulbright Program, the National Science Foundation, and the Wenner Gren Foundation for Anthropological Research. She also holds bachelor’s and master’s degrees in cultural anthropology from the University of Bologna, Italy.
Briana Vecchione is a technical researcher at Data & Society, where she investigates what happens when people turn to AI for mental, emotional, and behavioral health. Briana’s work focuses on developing user-centered methods and frameworks to inform the responsible design, evaluation, and regulation of AI in socioemotional contexts. More broadly, she is interested in algorithmic auditing and institutional accountability, and her previous work includes foundational work on dataset transparency, the history and social justice dimensions of audit methodologies, and empirical studies of audit tooling gaps and regulatory compliance. She also serves as an institutional partner for Brown University’s AI Research Institute on Interaction for AI Assistants (ARIA), where she leads participatory, empirical research to inform the landscape of AI care and support. She holds a PhD in information science from Cornell University, and her work has received support from the NSF, MacArthur Foundation, Mozilla Foundation, Google, Meta, IBM, and Microsoft.