Presentation
Perspectives on Ecological Validity and Real-Time, Multi-sensory Data for Human-AI Systems
DescriptionAs AI-enabled systems development reaches the limits of what can be done with historical data, synthetic data, and simulations, there is a growing urgency to improve human-AI-team performance andAI judgment via various strategies, including building world models, developing rich, multi-sensory, real-time sensing and perceptual systems for AI-enabled systems, and rigorous and comprehensive assessments of ecological validity. Critical, missing components in these discussions are the perspectives of the human factors specialists who observe, monitor, model, simulate, and design for human-AI systems performance. This panel presents perspectives on this topic from human factors researchers in the Augmented Cognition, Human-AI-Robot-Teaming, Human Performance Modeling, Perception and Performance, and Systems Development Technical Groups — i.e., researchers who will use ecologically valid assessments of Context of Use and Operational Design Domain, as well as multi-modal and multi-sensory, real-time sensing and perceptual data as inputs for their respective modeling, simulation, and systems development work.
Contributors
Event Type
TGPC Submission
TimeWednesday, October 21st4:30pm - 5:30pm PDT
Location