Projects

End-to-end UX research and design projects, from problem framing and prototype to evaluated design principles.

Click any method tag to filter
Trust CHI 2021

Does Transparency About Training Data Change How Users Trust AI?

Two studies, 44 participants. What happens when users can see what an AI was trained on?

Trust rose with good training data and fell with bad, consistently across all 4 AI scenarios and regardless of ML expertise.
Prototype Design Interview Study Iterative Design Thematic Analysis Mixed Methods
View project →
Style VL/HCC 2024

Does Explanation Presentation Format Affect How Users Assess and Understand AI?

Same information, two presentation formats. Does how AI explains itself change how well users can evaluate it?

Narrative format produced 66% high-accuracy critiques vs 42% for Q&A. Same content, presentation alone drove the gap.
Prototype Design Controlled Experiment Qualitative Coding Mixed Methods
View project →
Depth IUI 2026

How Much Detail Should a Training Data Explanation Include?

Conventional wisdom says simplify AI explanations. This study pushed back.

More detail produced more and better critiques. Users willingly accepted the extra cognitive load, contradicting the "simplify AI" assumption.
Prototype Design Factorial Design Cognitive Load Mixed Methods
View project →
AILS CHI 2025

How Do We Measure AI Literacy: Developing a Scale That Captures Both Knowledge and Confidence?

476 participants, two independent validation samples. How do you measure what someone actually knows about AI vs what they think they know?

High self-raters were the least accurate. Calibration mismatch confirmed across 476 participants in two independent samples.
Scale Development Survey Research Psychometric Validation Statistical Analysis
View project →