Is this project an undergraduate, graduate, or faculty project?
Undergraduate
Project Type
group
Campus
Daytona Beach
Authors' Class Standing
Kelly Nguyen, Senior Olivia Hartmann Kailey Hrbek Madeline Nees Gabrielle Roth Emily Silliman
Lead Presenter's Name
Kelly Nguyen
Lead Presenter's College
DB College of Arts and Sciences
Faculty Mentor Name
Barbara Chaparro
Abstract
Amphibious emergency support operations involve rapidly changing information, high stress, and significant cognitive demands, which can make decision-making and situation awareness more difficult for operators. When interfaces are poorly designed, they can contribute to issues such as alarm flooding, confusion from incomplete information, and delayed responses, all of which increase operational risk during time-critical disaster situations. This study explores whether using generative AI to assist with interface design will improve performance (output quality and effort) and usability compared to a manual sketch mock-up. Participants were asked to design a dashboard interface to support disaster relief operations following a Category 5 hurricane, where a Navy hospital ship provides offshore medical evacuation and logistical support. The interface needed to track injured civilians, monitor bed availability and medical supplies, display a common operational picture through a map, and summarize incoming communications in a chaotic and fast-changing environment. Participants were divided into two groups: a manual design group that relied on traditional resources and human-factors standards, and an AI-assisted group that used generative AI tools such as Figma Make, Claude, or Visily. Performance was evaluated using task completion time, perceived cognitive workload through NASA-TLX, usability ratings using the System Usability Scale (SUS), heuristic evaluations of the final interfaces, and participants’ feedback on which interface designs they preferred. Because analysis is still ongoing, we report preliminary results. These results suggest that participants who used generative AI produced interfaces with higher fidelity and greater usability than those in the manual group. These findings suggest that AI tools may help support more effective human-centered interface design in complex and time-critical operational environments.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Included in
Artificial Intelligence and Robotics Commons, Operational Research Commons, Systems Engineering Commons
Humans vs. AI: Comparing Approaches to Disaster Response Interface Design
Amphibious emergency support operations involve rapidly changing information, high stress, and significant cognitive demands, which can make decision-making and situation awareness more difficult for operators. When interfaces are poorly designed, they can contribute to issues such as alarm flooding, confusion from incomplete information, and delayed responses, all of which increase operational risk during time-critical disaster situations. This study explores whether using generative AI to assist with interface design will improve performance (output quality and effort) and usability compared to a manual sketch mock-up. Participants were asked to design a dashboard interface to support disaster relief operations following a Category 5 hurricane, where a Navy hospital ship provides offshore medical evacuation and logistical support. The interface needed to track injured civilians, monitor bed availability and medical supplies, display a common operational picture through a map, and summarize incoming communications in a chaotic and fast-changing environment. Participants were divided into two groups: a manual design group that relied on traditional resources and human-factors standards, and an AI-assisted group that used generative AI tools such as Figma Make, Claude, or Visily. Performance was evaluated using task completion time, perceived cognitive workload through NASA-TLX, usability ratings using the System Usability Scale (SUS), heuristic evaluations of the final interfaces, and participants’ feedback on which interface designs they preferred. Because analysis is still ongoing, we report preliminary results. These results suggest that participants who used generative AI produced interfaces with higher fidelity and greater usability than those in the manual group. These findings suggest that AI tools may help support more effective human-centered interface design in complex and time-critical operational environments.