There are two tech catchphrases of the last few months, one closely linked to the other.
The first concerns the increasingly futuristic arrivals of artificial intelligence, especially generative intelligence, i.e. producer of new content. We were all amazed, for example, when we saw the quality of the video clips produced by Sora, OpenAI software capable of transforming a text command into an extremely realistic video lasting one minute.
The second is an almost direct consequence of the first, namely an increasingly urgent need to regulate artificial intelligence. In this sense something, or rather a lot, at least in Europe is moving: On Wednesday 13 March the European Parliament gave its approval to the AI Act. Let’s take a closer look at what happened and remember what the Artificial Intelligence Act is.
Ai ACT: green light from the European Parliament
The European Parliament gave the AI Act the green light with 523 votes in favour, 46 against and 49 abstentions. Europe will therefore have its own regulation on artificial intelligence.
It will apply to anyone, public or private, who produces artificial intelligence software intended for the Community market or whose use concerns people located in the European Union.
Only artificial intelligence systems for military, defense or national security purposes, for scientific research purposes, are excluded. In addition to those released under free and open source licenses.
L’iter
The European regulation on artificial intelligence was born in April 2021 following a proposal from the European Commission.
After the favorable opinions of the Council and the EU Parliament, we arrived at the painful session of last December 9th, which we gave you an account of in an article. After 22 hours of disagreements, the political agreement of the trilogue arrived, i.e. the joint meeting of representatives of the Council, Parliament and Commission.
Until the European Parliament approves the AI Act. And now? The regulation will come into force in stages. The first step will be, within six months, the identification of prohibited practices, but full implementation will have to wait two years.
The AI Act at a glance
The AI Act, as reported on the official website, is “a European law proposal on artificial intelligence, the world’s first AI law from a major regulatory body. The law assigns AI applications to three risk categories.”
The regulation works by following three risk categories: the higher risk the AI systems are, the more stringent the measures of the provision are.
There are some AI systems that are completely off limits. Among these, predictive policing systems: law enforcement agencies will be able to use biometric recognition systems only in exceptional cases, such as an imminent terrorist threat, the search for victims of serious crimes or the prosecution of serious crimes.
Facial recognition is also prohibited, except “in the targeted search of a person convicted or suspected of having committed a serious crime”.
Also prohibited are the recognition of emotions in schools and workplaces, social scoring and techniques that aim to manipulate human behavior.
The sanctions
Companies that do not comply with what the European regulation on artificial intelligence will impose, from the AI Act, may incur different sanctions depending on the seriousness of the crime.
Fines will start at 35 million euros or 7% of global annual turnover (the higher of the two values) for violations of prohibited AI applications. Or 15 million euros or 3% of turnover for other violations, and 7.5 million euros or 1.5% for dissemination of incorrect information.
The statements
Thierry Breton, European Commissioner for the Internal Market, commented with some enthusiasm on the European Parliament’s approval of the AI Act.
In a post on X, Breton wrote: “Democracy: 1️-Lobby: 0️. I welcome the European Parliament’s extraordinary support for our #AIAct, the world’s first comprehensive and binding rules for trustworthy AI.
Europe is now a global regulator of standards in the field of artificial intelligence. We are regulating as little as possible, but as much as necessary!”
And PD MEP Brando Benifei, co-rapporteur of the European Parliament’s Internal Market Committee, added: “We protect the rights of workers and citizens. We will now have to accompany companies to comply with the rules before they come into force. We have managed to put humans and European values at the center of AI development.”
Stay updated by following us on Google News!
Follow!