A paper I co-authored on how much an AI companion agent should disclose about itself, and what that does to how people perceive, trust, and engage with it. I built the full-stack web app that ran the study and logged the interactions, on a data pipeline that met the university's research protocol; then we measured perception across a few disclosure levels with validated surveys and analyzed the results in R. It turns out how much an agent 'opens up' really does move trust and likability.
In the Ruiz HCI Lab at UF I build the software behind human-subjects research, eye-tracking, augmented reality, and how people trust and use AI. That runs from real-time gaze pipelines to full study apps with IRB-compliant data collection, and into the analysis afterward. I like research questions you can only really answer by building the thing a person then sits down and uses.
Browse the full collection →