THIS IS ABSOLUTELY PROHIBITED

The bill lists some functions that just can’t be done with AI.

This is because Europe fears a very high risk to our security.

– First of all when governments transform AI to create a social profile. The so-called “social score”.

As I explained above:

“A way of obtaining an assessment of the behavior of individuals from the authorities (such as the state police). A social scoring.”

– According to the use of remote biometric identification systems made in real time.

They are especially prohibited in spaces accessible to the public for the purposes of law enforcement (with certain exceptions).

– Third, subliminal techniques aimed at madly altering a person’s behavior in a harmful way.

Yes, AI can do this too. Incredible right?

Then…

WHEN THERE IS A HIGH RISK FOR PEOPLE

The regulation proposed by the European Parliament establishes this specific category for high-risk AI systems.

That is, it is the technologies that could cause harm: their use is okay only in the presence of a security check.

Among these are:

– Credit scoring; that is, how much a person is financially at risk. Maybe if he has more than one mortgage and is able to pay it or not. Or if it has defaulted on some loan in the past.

– AI systems related to essential public infrastructures: such as justice and social security.

– And medical devices and other regulated devices, and transportation systems.

All these examples fall into two categories into which they have been divided. One of the two is very important.

I’m talking about biometric identification systems both in real-time and in retrospect.

“Biometric identification is a system based on a mathematical algorithm that does an apparently cool thing: it identifies a person based on some of his main physiological and behavioral characteristics.”

Like Shazam for music. Don’t tell me you don’t know what Shazam is!

You know, that application that many have on their smartphone: if you are at a Café and on the radio you hear a song you like but you don’t know, you activate that app and she tells you the title of the song because she can recognize it. She has them all filed in an archive.

It is a bit more complicated but this is how the identification of artificial intelligence would work. For example, through the CCTV cameras of a building. It just goes against people’s privacy.

And above all the fundamental rights of individuals.

The last thing about high-risk systems is that:

The proposed regulation for AI provides for a series of checks to be carried out on these high-risk algorithms by the provider. If all this becomes law it will be an obligation.