The Community 's "detection of lies" in the EU is misleading


Calling the new EU border control AI as a "lie detector" is like asking Brexit for a little clash between friends.

The lowest level is that the EU is testing a pilot project for international airports that includes a mechanical lie detector. CNN broke history last week his article "Passengers to face AI dummy detector tests at EU airports."

According to the report (and the project website), the EU is testing a pilot project involving AI using an avatar to ask people questions. It is supposed that this AI-powered construct knows if a person is true when it answers and they mean those that it determines that it is not. When someone is a liar, people are called upon to explore further.

Depending on your views on privacy and migration, this is either the music in your ears or the beginning of a dying future straight from a nightmare Orwell. You make a mistake in every way.

For example, if you think "we could just ask everyone" are you a terrorist? "and make the EU safer for everyone" then you probably suppose there is such as a lie detector AI. There is no. Do not worry, this is a common mistake.

Like all lie detectors, AI-based solutions are based on hit-or-miss cases with what the government considers a tolerable wrong level. A polygraph, for example, is reported to have an accuracy rating of 75 to 90%. He does not detect lies, he detects the subtle slogans associated with the lie. Exactly, AI does not detect lies. It detects biometric indicators related to the lie – with approximately 76% success, allegedly.

On the other hand, if you think this is the most flagrant violation of human rights at an airport since the days when TSA agents in the US were instructed to record Muslims, you were wrong. Mostly, anyway: it goes far beyond just frenetic aliens at airports.

The "lie detector" is just the tip of the iceberg. What you need to know is that in this case the "lie detection" is short for "data collection".

The new EU airport plan is not called the "operating alarms detector". It is called "iBorderCtrl". And "detecting lies" is just a part of what it does. Here is an image from the project site:

Credit: iBorderCtrl

First, start with the Lie Detection Unit: Automated Tracking Detection System (ADDS) *. Star is from iBorderCtrl, TNW did not add it. Before you read about what ADDS does, the people who build it are obliged to highlight the following:

* ADDS relies on previous developments, the so-called Silent Talker [1][2]. The project has adopted this technology and is well aware of the controversy surrounding it. This tool has been subjected to the following steps: A scientific background is achieved when a research starts from a place of ignorance and follows the scientific method to dissolve this ignorance. As AI scientists, a specific research question is "There are non-verbal indicators of cheating behavior," and the experiments that are being conducted collect data to support or refute the underlying assumptions.

Basically, the project recognizes, but is not interested in, a dispute over the use of AI for "detecting lies".

The rest of the explanation for the ADDS is full of the old "non-verbal index" argument. It is academically interesting, but it is certainly not the irregular development we needed to make the AI ​​lie detector. Therefore, shame in any government to consider it at this point.

But ADDS is the silly insignificant part of iBorderCTRL. It is the fake monster that distracts you from the cold, computed data collection that really happens. Basically, ADDS is a low-hanging fruit for experts in the spot and stoke talks about the possibilities AI really has. Meanwhile, if I can paraphrase The Ragin Cajun: "It's the data, stupid."

In fact, the rest of the iBorderCtrl suite is scary. ADDS closely resembles Cambridge Analytica's "personality profile" is used to persuade people to give up their data. After all, why did not you agree to take a lie detector if you have nothing to hide?

The answer is that you have a lot to hide if you are lying or not.

Here's what iBorderCtrl does:

  • Gets a full face profile using videos and photos
  • Scraps and scours all your social media accounts
  • Analysis of documents and signatures
  • Creates and stores your digital voice printing
  • Risk assessment based on aggregated data
  • He's looking for hidden people

Let's get the last one for quick scrutiny. Searches for hidden people? At an airport? Unlikely. This, according to the bibliography of the project, is looking for people who are hiding. But it is painfully obvious that this is designed for cross-border borders where mobile agents are watching or looking for people outside, not at airports. Still, it's part of the "lie detector" airport that you listen to so much.

Simply put: iBorderCtrl is a complete data collection and tracking package that seems to be more powerful than a Trump put into use at the US / Mexico border.

The amount of information that the EU could get from it – everything ranging from a person's sexual identity to religious and political beliefs – would allow him to target people with laser precision.

But this is just a problem if you do not believe in biased data, corrupt government, or promises AI underground.

IBorderCtrl did not immediately send TNW request for comments.

Read below:

Killing a suffragette in Red Dead Redemption 2 tells you more about Rockstar