This 1 November 2018 video says about itself:
Translated from Dutch NOS TV today:
The European Commission is investing 4.5 million euros in a lie detector that is supposed to detect illegal immigrants at the border with artificial intelligence. The computer system, iBorderCtrl, would consist of an animated border guard that asks questions to travelers. The system would then analyze their “micro-facial expressions” to determine if they are lying about their travel plans.
This reminds me of the Palantir Big Brother artificial intelligence software (ab)used by authorities in New Orleans and elsewhere in the USA.
Maybe that European Union border software might ‘recognize’ me, or you, as the late Osama bin Laden, or at least his ghost.
The intention is to experiment with the system at border crossings in Hungary,
Greece and Latvia.
The European Commission speaks of a “smart lie detection system”. The AI lie detector should make it easier for border guards to recognize illegal migrants “and thus contribute to the prevention of crime and terrorism“.
Oh yeah, the myth that people crossing borders without documents are supposedly more criminal and more terrorist than ‘autochthonous’ people. The European Commission parroting Donald Trump’s xenophobia on refugees from Honduras and other Central American countries.
The company behind the lie detector believes that the system achieves a success rate of 85 percent in the laboratory. But critics have strong doubts and are worried about the plan.
According to the science magazine New Scientist, iBorderCtrl has been tested on just 30 people. In the Dutch Volkskrant daily, a scientist says that there is no evidence for the popular assumption that liars are stressed and that this stress translates into fidgeting or a trembling corner of the mouth. Nor would any reputable lie expert cooperate with the project.