Facial Lie Detection Tech Being Developed for Airports, Border Crossings
- By Cindy Horbrook
- Mar 09, 2011
Could facial lie detector screening be on its way to an airport or border crossing near you?
Researchers at the University of Arizona hope so. The AVATAR (Automated Virtual Agent for Truth Assessments in Real-Time) kiosk project is currently under development at the university as part of the Department of Homeland Security-funded, 14-university BORDERS consortium.
The kiosk is an interactive screening technology designed to be on the front lines of border crossings and airports. Individuals would approach the kiosk, scan their identification and answer a few questions. In the mean time, the kiosk has used non-invasive artificial intelligence and sensor technologies to gauge suspicious behavior. The current version of the kiosk assesses cues through sensors in body movement, vocalics, pupillometry and eye tracking.
Lead developer Doug Derrick has been studying how to detect deception in interpersonal communication for about three years as part of his PhD dissertation.
“The goal was to create something that could automate that process, be noninvasive, and provide additional data to agent that they couldn’t provide on their own,” he said.
Derrick stressed the kiosk is designed to be a system support for the human and would not have the final word.
“The point of the system is not to replace the human,” he said.
Stephen Russell, chairman and founder of 3VR Security, and editor-in-chief of InHardFocus, thinks it’s too early to know whether such a kiosk could make its way into airports, but has some doubts. He said issues of cost, training, legality and an inevitable public backlash will likely keep the technology, like other lie detectors before it, out of airports.
“In 2005, we saw similar stories about ‘voice-based lie detection’ being deployed at airports in Israel and Russia, but issues of accuracy, implementation and legality plagued the rollouts,” he said. “Also, despite being quite advanced, it just couldn't match the ingenuity and flexibility of highly trained humans. I suspect ‘facial lie detection’ will encounter the same issues.”
Derrick said the kiosk has so far proved to be far more accurate than a human at lie detection.
“A wide body of research shows humans are only accurate 54 percent of the time at detecting deception. People are really poor lie detectors,” he said. “In the studies we’ve run in the lab, we were accurate about 89 percent of the time.”
The kiosk prototype was put to the test last fall in Warsaw, Poland, when researchers ran an experiment with a participating group of European Union border guards. Half of the guards constructed mock bombs and packed them in suitcases, then tried to get through the AVATAR kiosk. Derrick said the kiosk detected all of the mock bombers, but captured some false positives as well.
It’s probably still be three to five more years before AVATAR will be ready to go out into the field. One thing researchers are still working on is evaluating cultural differences.
“Deceptive behavior can vary by culture, so we’re looking at how culture could impact,” Derrick said.
Whether or not the kiosk makes it into airports, Stephen Russell from 3VR Security agrees something needs to be done about airport screening in the U.S.
“Today the bar is pretty low, at least, as far as airport screening procedures in the U.S. are concerned. TSA screening's most visible ‘successes’ may be needless confiscation of close to five million bottles of shampoo, tweezers and tubes of toothpaste. If ‘facial lie detection’ could replace these procedures, accurately and efficiently separating the good guys from the bad guys at check-in, perhaps we could eliminate the awful and unpopular privacy-invading scanner checkpoints we all endure today.”