Border Control Agencies May One Day Use AI to Detect Travelers’ Lies

Aaron Elkins
Aaron Elkins / Aaron Elkins
facebooktwitterreddit

Border control agencies are already using self-service kiosks to manage the crowds of international travelers entering their countries, but a high-tech type of kiosk in development can do more than just scan passports. The AVATAR—which stands for Automated Virtual Agent for Truth Assessments in Real-Time—can detect travelers trying to lie their way through customs, according to Vocativ.

The self-service kiosks, created by the National Center for Border Security and Immigration at the University of Arizona in partnership with the Department of Homeland Security [PDF], scan travelers' passports and ask the kinds of questions posed by human agents, such as “Do you have any fruits or vegetables?” Sensors can identify body cues like facial expression, vocal tics, pupil dilation—and even cues that human agents can’t see, like cardiorespiratory data—which could indicate that the person is lying and should be subject to additional screening. They can even see that you’re curling your toes, according to a press statement from AVATAR researcher Aaron Elkins of San Diego State University, a professor who studies deception.

The kiosks can be programmed to display several virtual agents, choosing from a woman or a man and a stern or a friendly face. They can be configured to interview in several languages.

AVATAR has been tested in a number of experiments in the European Union and North America, including a pilot program in Nogales, Arizona, in which it screened passengers in the Trusted Traveler program for suspicious or unusual behavior.

"AVATAR has been tested in labs, in airports, and at border crossing stations," Elkins said in the press release. "The system is fully ready for implementation to help stem the flow of contraband, thwart fleeing criminals, and detect potential terrorists and many other applications in the effort to secure international borders." However, not all machines built to detect lies are accurate—polygraph tests are largely useless at sussing out dishonesty, according to many psychologists—so there are plenty of reasons to proceed cautiously with this kind of technology.

[h/t Vocativ]