Designing Adaptive AI
for
Human Agency

I am interested in how intelligent systems can become perceptive, adaptive, and collaborative partners in everyday human experience. As AI moves beyond screen-based interaction into wearable, embodied, and ambient environments, it can increasingly sense our physiological states, behaviors, activities, and surroundings—and use these signals to decide when and how to intervene.

​

My research asks how such systems should participate in the loop between human perception, interpretation, decision, and action. I build multimodal and embodied AI systems that sense human and environmental context, construct dynamic models of users' cognitive and bodily states, and adapt their behavior through spatial, auditory, and tangible interactions. Rather than treating adaptation simply as personalization, I am particularly interested in the intelligence required to decide not only how to act, but when to wait, remain uncertain, or defer to the user.

​

Ultimately, I want to understand how humans and AI can form adaptive joint cognitive systems without diminishing human agency. As AI becomes increasingly capable of interpreting our bodies, cognition, and behavior, how can we design interactions that help people better understand themselves, retain authorship over their decisions and actions, and actively negotiate the models that AI constructs about them? My long-term goal is to develop human–AI systems that continuously learn, adapt, and evolve together with the people they support.