Three airports within the EU are starting to use lie detectors to check passengers wishing to enter Greece, Hungary and Latvia.
But the method is controversial and does not guarantee catching people who are lying.
Soon, non-EU air passengers who want to visit Greece, Hungary and Latvia may be met with a lie detector based on AI, artificial intelligence. Passengers will be required to stand in front of a webcam and a computer-animated border control avatar.
The avatar asks a number of questions to the person who wants to cross the border and reads the passenger's facial expressions. The questions may include name, age, purpose of the trip or who paid for the trip.
Anyone who acts nervously or otherwise causes the avatar to interpret the passenger's facial expressions as lying will be forwarded to a human border controller, reports CNN.
Problems with the technology
But the system, which goes by the name iBorderCtrl, has received massive criticism. Previous similar systems have had problems reading female and dark-skinned faces. The AI avatar that will be tested in Greece, Hungary and Latvia has been tested on 32 people and the manufacturers believe that it should be able to get 85 percent correct information.
“Traditional lie detectors have a problematic history of accusing innocent people of being criminals. There is no evidence that an AI can handle that – especially not a system that has been tested on 32 people,” says Frederike Kaltheuner, data program manager at the organization Privacy International.
He believes that innocent people suddenly find themselves in a situation where technology believes they are guilty of a crime and they themselves must prove that they are innocent.
Initially, only passengers who give their consent via an online document will come into contact with the avatar. Testing with iBorderCtrl will start in November, securitytoday.com reports.
Source: aftonbladet.se
