Beyond Human Actors: Leveraging AI for Enhanced Public Safety XR Training
Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 2025
I was an undergraduate researcher at the NeuroErgonomics Lab, under Dr. Ronak Mohanty and Dr. Ranjana Mehta, from summer to winter 2024. We examined the efficacy of embodied AI NPCs at replacing human play-actors in law enforcement deescalation training scenarios. I mainly helped implement the AI NPCs and develop a prompting framework for them, as well as set up the data collection and analysis pipeline for implicit bias, eye tracking, and heart rate data. I ran many participant studies with UW students, and later UWPD officers, and helped train two new undergrads to run studies in the fall.
We found that officers participating in our trials decreased their implicit bias scores by 50%, for race and gender.