Belgien, um KI zu nutzen, um Kriminelle mit dem ersten engagierten Minister zu finden

The first use of AI by the Belgian government is basically to use it to spy on it’s people’s past and present actions to punish them. Jus wow.
They should instead use AI for what it is meant and advertised for: increase productivity and reduce unnecessary government spending.

Great, now we can find criminals faster, then try them in court two years later, threaten them with non effective punishment, and release them on a technicality!

But seriously, I wonder what the ‚use by law enforcement authorities in specific cases‘ means. It’s good that this can’t be used to keep track of protesters for example. But I guess facial recognition is useful in case of a manhunt for a big criminal and also to monitor common litter spots for example. But which of these would be legal?

You know, instead of going weird ways with AI and everything, you could also just try hiring enough police so that they may answer their emails in less than 2 months… Bonus if it means they can go to accident locations on the day they are called.

We are doomed! Religious Flemish fruitcakes are in charge.

I’m not against the use of AI and facial recognition to fight crime. But in this case it feels like giving a loaded automatic rifle to a five year old with muscle spasms. 

I have zero confidence in the Belgian police being able to do basic police work as it is, a tool like this won’t help one bit.

We need police reform, better trained police officers, more police presence, and once they prove they can actually do their job in an efficient and trustworthy way – THEN let’s discuss adding powerful privacy infringing tools. 

Right now I’m certain that such a tool will only be used to decrease police presence and make them even slower to respond. It wouldn’t even surprise me if they’d use it more to identify people putting their bags out on the wrong day than to identify bike thieves and robbers. 

>In navolging van de Artificial Intelligence Act die recent werd gestemd in het Europees Parlement, maken we werk van een concreet beleid rond het inzetten van technologieën die onder toepassing van deze AI Act vallen. We maken dan ook in samenwerking met de inlichtingendiensten en veiligheidsdiensten – binnen een strikt en precies bepaald wettelijk kader- proeftuinen voor nieuwe operationele toepassingen en nieuwe technologieën mogelijk . Daarbij kan bijvoorbeeld gedacht worden aan de inzet van gezichtsherkenningstechnologie voor de opsporing van veroordeelden en verdachten. We verruimen de camera-wetgeving om meer toepassingen mogelijk te maken van (slimme) camerabewaking. Ook AI wordt verder ingezet om de politie van bepaalde administratieve taken te ontlasten, zoals het invoeren van ‘speech-to-text’ principe. Tijdens het onderzoek en de mogelijke uitvoering/invoering van deze nieuwe technologieën maken we altijd de afweging tussen het recht op veiligheid (wat een fundamenteel mensenrecht is) en het recht op privacy.

This is from the „regeerakkoord“. In the AI Act it says that the following is prohibited:

* **compiling facial recognition databases** by untargeted scraping of facial images from the internet or CCTV footage.

Which means that police forces could run AI on targeted or known footage of perpetrators. Or, use AI on random and public cameras without storing anything they find. Something like that. But how do you make sure police forces adhere to those rules? What will be the controls? And that last sentence scares me as well (right of safety vs right of privacy, where do you draw the line?)

Next up: social credit system.
Fuck this.