Professor McCarley has just received a $235,109 grant from the US Department of Homeland Security to develop guidelines for ensuring security software is designed in a way to optimise human screeners’ performance.
“When you go through the airport you pass your bags through an X-ray scanner, and maybe have to walk through a whole-body X-ray scanner yourself, so that screeners can search for weapons or other contraband,” Professor McCarley, based in the School of Psychology, said.
“Computerised search aids are designed to pinpoint or highlight suspicious objects in the X-rays to make sure the human screener doesn’t miss anything,” he said.
“But just like a human screener, computerised aids make mistakes – they don’t always spot targets that are there and they sometimes produce false alarms.
“The consequence is that the human screener often stops using the aid, even when it could be helpful, so our goal is to design the interface and behaviour of the computer aid to help the user trust it more.”
As part of the year-long project, Professor McCarley will investigate how errors committed by search aids influence user behaviour, and establish principles to minimise the psychological costs of the computer’s errors.
“People tend to distrust the computer aid more after it commits a false alarm than after it misses a target,” he said.
“The designer can modify the system to reduce false alarms but that will increase the number of targets that are missed.
“Therefore, the design we choose needs to balance errors in a way that encourages the user to trust the software appropriately.”
Professor McCarley said it was important to make global travel as safe and expedient as possible.
“We want people to travel safely, and we want to get them through the security lines as quickly as possible in a way that minimises stress for the screeners and passengers.
“The funding from Homeland Security will help us figure out how best to do that.”