From Alexa and self-driving cars to job applicant screening processes, artificial intelligence is fast becoming the norm in business. But it also could start playing far bigger roles in security, helping law enforcement and other protective agents figure out who's up to no good. As Fredrick Kunkle of The Washington Post reports, there's now an AI-based kiosk designed to detect whether travelers are fibbing.
Designed by Aaron Elkins, assistant professor of the Fowler College of Business Administration at San Diego State University, the new AI lie detector goes by the name Automated Virtual Agent for Truth Assessments in Real Time, or AVATAR for short. Once you've scanned your ID or passport, the kiosk asks you a bunch of questions. The inquiries are a good mix of inquiries you could practice (e.g., when were you born) and questions that might throw you if you're faking it (e.g., describe what you did today). You can see some of the process in the video below:
If everything goes well, security personnel should let you go on your way. If the results suggest you're being dishonest, security personnel might detain you for questioning or a search.
How the AI sorts liars from truth tellers
As you answer questions from AVATAR, the system uses sensors to gather data your body gives off. More specifically, the system looks at factors like voice (tone, pronoun use, etc.), pupil dilation and eye movement, facial expression (e.g., engagement of muscles around the corners of the eyes and mouth in a Duchenne smile) and posture. The theory is that it takes less effort to tell the truth than to maintain a façade. You subconsciously reveal that effort through physical cues, many of which researchers are still studying and pinning down. The AI is a big step forward from traditional polygraphs, which aren't practical for general, large-scale screening, use more limited physiological data (e.g, heart rate) and generally aren't considered very reliable.
Where you might see AVATAR at work
In theory, AVATAR could become a widely applied staple in local law enforcement agencies around the world, helping police sort out a variety of conflicts. But its main intent is for border security checkpoints and airports. These facilities are of concern in part because of the high traffic they receive. But they are also worry points because of the current worldwide focus on terrorism. Although these types of attacks can come from many different individuals or groups and can be domestic or foreign in nature, the increasing activity of the Islamic State of Iraq and Syria (ISIS) has been particularly alarming for leaders around the globe. Attacks have led U.S. President Donald Trump, for instance, to call for a controversial travel ban against travelers from six majority-Muslim countries. AVATAR might one day help screen out individuals associated with ISIS or similar groups.
Right now, AVATAR is still in its infant stages. It's only collecting research data at border crossings in Mexico and Romania. But even at this point, it's a beautiful demonstration of how science and technology can blend toward a practical social good.