University of Michigan - School of Information
Improving an AI Chatbot for healthcare
Summary
This internship centered on a cross-cultural study of parent-child communication and emotional expression in the context of emerging AI technologies. The research examined how parents in the United States and South Korea (with children aged 8–12) communicate with their children and how these dynamics influence the design of child-facing AI systems.
As a UX Research Intern, I contributed across the end-to-end research cycle, from participant recruitment to qualitative analysis. My work emphasized ethical recruitment practices, semi-structured interview support, and thematic analysis that informed recommendations for the design of AI chatbots supporting family communication.
Research Approach
Interviews
Conducted semi-structured interviews exploring:
Parent–child communication practices
Children’s emotional expression habits
Parents’ perceptions of AI in family life
Strategies to strengthen family interactions
Reflections on the persona of an AI system (ChaCha)
Supported facilitation, captured detailed notes, and cross-validated with transcripts for accuracy.
Analysis
We applied thematic analysis using Atlas.ti and Miro.
Codes captured themes such as:
Emotional communication practices
Opinions about AI
Children’s communication habits
Parental strategies for connection
Feedback on ChaCha’s persona
Synthesized insights collaboratively, validating them across cultural contexts.
Affinity Mapping overview
Closer view of affinity mapping
Impact
Designed recruitment materials that secured 12+ parent participants per country.
Supported interview facilitation and note validation, improving data reliability.
Coded and analyzed data to surface recurring cross-cultural themes informing AI chatbot design.
Collaborated with researchers in the US and Korea, contributing to a nuanced human-centered AI research output.
Takeaways
Learned the importance of culturally sensitive UX research in cross-national contexts.
Gained hands-on experience with thematic coding and collaborative analysis tools (Atlas.ti, Miro).
Strengthened skills in aligning diverse stakeholders to ensure clear, actionable design recommendations.
Developed awareness of how parental perspectives on AI influence the design of child-facing AI technologies.
Interview with a parent for feedback on the AI chatbot