Human-centered AI
My research asks how AI systems can be designed so people can understand them, trust them appropriately, and use them to improve individual, organizational, and community outcomes. I bring an HCI perspective to this question, connecting the design of intelligent systems with the experiences and needs of the people who use and are affected by them.
My program develops through three connected directions: foundational work on conversational explanation, translation into healthcare, and accountable systems designed with communities. Together, they connect interaction design with clinical and community contexts, while creating opportunities for students to contribute to research, publications, and public engagement.
Research direction 1
I investigate how explanations can become part of an interaction rather than a static account of an algorithm. My work connects transparency, user control, and conversational interfaces so people can question AI outputs, interpret recommendations, and apply explanations in their own decisions. Through studies of social recommender systems and interactive explanation interfaces, I develop design principles for human-AI collaboration that account for differences in knowledge, goals, and AI literacy.
2020 · User Modeling and User-Adapted Interaction
A controlled study with 50 participants examines how user control and visual explanations work together in an interactive social recommender system.
2024 · Disruptive Technologies, Tech Ethics and Artificial Intelligence
An LLM-based recommendation architecture and an 11-participant pilot explore context-aware explanations and the information users want from a social recommender.
2022 · xxAI – Beyond Explainable Artificial Intelligence · Springer
This chapter argues that useful AI explanations must address intelligibility, trustworthiness, and practical use alongside computational accuracy.
Research direction 2
I translate conversational and explainable AI into healthcare systems that make information more accessible, contextual, and actionable. This direction spans nutrition and maternal health, caregiver information seeking, healthcare navigation, early autism screening, and clinical communication. I combine system design with empirical evaluation to understand how AI can support patients, caregivers, and clinicians. My CAREER program builds on this work to advance nutritional health and dietary counseling through human-AI collaboration.
2023 · ISCRAM
An early NutritionBot study co-designs pregnancy nutrition guidance with healthcare experts and develops a persona for users from underserved populations.
2025 · UMAP / FairUMAP
A study with six caregivers and five healthcare providers examines a retrieval-augmented chatbot and differing expectations for the depth and presentation of health information.
2026 · International Conference on Digital Government Research (dg.o)
A multimodal framework uses caregiver videos to surface interpretable behavioral evidence for early autism screening, while showing uneven performance across behavior categories.
Research direction 3
I study how communities can help shape the intelligent systems that affect them. Working across civic, environmental, educational, and emergency-management contexts, I use participatory research and co-design to connect technical choices with community knowledge, institutional responsibilities, and local needs. A central focus is accountable AI designed with Tribal Nations, including disaster-management systems, collaborative governance, and the preservation of Indigenous culture and sovereignty. Related work examines citizen science, mutual aid, and collective responses to public crises.
2024 · Digital Government: Research and Practice
Participatory studies with Tribal Nations examine how disaster-reporting chatbots and drones can support emergency management while respecting sovereignty.
2026 · International Conference on Digital Government Research (dg.o)
Survey data from 146 participants representing 63 Tribal Nations connect perceived AI accountability most strongly with transparency and perceived benefits.
2025 · Proceedings of the ACM on Human-Computer Interaction (CSCW)
Participatory design with citizen-science groups identifies infrastructure needs for water-quality monitoring, including data standardization and resource sharing.