Home | Ai For The Well Bein

Behavioral Intelligence Lab @HKUST
行为智能实验室

The Behavioral Intelligence Lab works toward a human-centered future where AI truly serves human wellbeing. Many of the hardest problems in today's wellbeing crisis are behavioral: people who need care don't seek it, don't stay with it, and are rarely reached by the services that could catch problems early and intervene in time. At bottom, these are problems of measurement, demand, matching, and engagement — and they are the problems we work on with AI.

We develop the science and methods for understanding how humans and AI agents behave and make decisions. We use that understanding to match people with the care, services, and products they need, to keep them engaged, and to turn behavioral data into instruments of care — so that technologies built to raise our productivity can also learn to care for human minds.

Our members are from multidisciplinary backgrounds, including computer science, economics, quantitative marketing, cognitive science, and psycholinguistics. We work to be a place where these different ways of thinking sharpen one another.

Research Theme

1. Computational Behavioral Science of AI and Human

We model, analyze, and explain the behavior of human and AI agents, as service providers or consumers, from individuals to markets. Recent works:

Behavioral alignment and deviation between humans and AI agents
Optimizing AI memory for user engagement
Simulation and evaluation of synthetic patients

2. AI for Mental and Social Wellbeing

We hope our research can contribute to a human-centered future in which AI technology serves human wellbeing, especially from mental and social perspectives. Recent works:

Building multi-agent screening systems for mental wellbeing
Designing AI-assisted interventions for mental wellbeing
Human–AI affective relationships for social wellbeing

Core Faculty Members

To make our research truly interdisciplinary, researchers from various fields collaborate closely together. Lab students can benefit from co-supervision by faculty members with diverse expertise.