Researchers Launch National AI Institute Focused on Mental Health at Brown University

Brown University hosted researchers from across the country to formally launch the new National Science Foundation funded AI Research Institute on Interaction for AI Assistants, known as ARIA. Supported by a five year, 20 million dollar NSF grant, ARIA aims to advance a new generation of AI systems capable of trustworthy, sensitive and context aware interactions with people, with a primary research focus on mental and behavioral health.

The ARIA initiative is led by Ellie Pavlick, associate professor of computer science at Brown, in collaboration with experts in computer science, psychology, neuroscience and related fields from multiple institutions. Pavlick noted that the kickoff meeting allowed researchers to identify key interdisciplinary research themes that will drive the institute’s work. The goal is to address major open questions in areas such as interpretability, adaptability, participatory design and the development of evaluation tools for trustworthy AI in mental health settings.

During the two day convening, the ARIA team held internal brainstorming sessions and conducted a public event to share the institute’s early plans with the broader research community. The public program featured a keynote talk by Yale psychologist Julian Jara Ettinger, who discussed how human social intelligence might inform AI systems that must understand user intent and emotional context. A response discussion was led by Brown clinical psychologist Nicole Nugent, who emphasized the importance of ARIA’s interdisciplinary framework.

Researchers from partner institutions, including Colby College, Dartmouth College, Carnegie Mellon University, the University of California San Diego and the University of New Mexico, presented short overviews of proposed research themes. These included efforts to create more interpretable AI systems, understand how AI can adapt to diverse users, and develop stakeholder informed design processes. The team also explored how to measure trustworthiness in AI for mental health, with early discussions comparing the goal to a Consumer Reports style evaluation system.

Brown Provost Francis J. Doyle highlighted the university’s strong culture of collaboration and its ability to serve as a central hub for scientific innovation. He noted that institutes like ARIA require close partnerships between researchers, practitioners and community stakeholders, and that Brown is well positioned to support this work over the next five years.