AI Act, what does the EU agreement on artificial intelligence provide

A historic achievement.
The Parliament and the Council of the EU have reached an agreement to approve the AI Act, the first law in the world to regulate the development and use of artificial intelligence systems.
After three days of intense negotiations, in the night between Friday 8 December and Saturday 9 December, political agreement was reached on a legislative text that was, to say the least, "revolutionary", as it aimed to indicate the permitted and prohibited uses.
to protect the privacy and other rights of European citizens.
At the time of the negotiations, Parliament and the European Council found themselves on clearly different positions – wanting to simplify – Parliament demanded more restrictive rules, while the Governments asked for greater margins of maneuver in the fields of security, the economy and in the industrial sector.
After reaching the political agreement, the text will now have to be refined by the technicians responsible for writing down the definitive version of the law, which will then have to be approved by the Parliament and the Council of the European Union.
The regulation will be fully applied two years after it comes into force, but what exactly does it involve? Below is everything you need to know about it.
read also 6 AI shares to be held until 2025 according to Morgan Stanley AI Act, What the rules of the agreement on Artificial Intelligence provide Compared to the community proposal, the AI Act provides enforcement powers at European level and its objective is to ensure that artificial intelligence protects: fundamental rights; democracy; the rule of law; environmental sustainability.
As regards biometric identification (RBI), it was agreed to grant permission to law enforcement agencies to use RBI but providing for a series of safeguards and restricted exceptions for its use, subject to judicial authorization and only for lists of strictly defined crimes.
In a note, Parliament explained that "post-remote" RBI would only be used for: the targeted search for a person convicted or suspected of having committed a serious crime such as terrorism, human trafficking, murder, rape; the search for victims of kidnapping, trafficking, sexual exploitation; the prevention of a specific and current terrorist threat.
As regards "generative" artificial intelligence – i.e.
the use of AI for the creation of new content, such as text, images, music, audio and video – the compromise involves a "two-speed approach": rules will be imposed to everyone to guarantee the quality of the data used in the development of the algorithms and to verify that they do not violate copyright; while developers will have to ensure that the content produced is clearly identified as "artificial".
Finally, the AI Act provides measures to support innovation and a sanctions regime, with fines ranging from 35 million euros or 7% of global turnover to 7.5 million or 1.5% of turnover, depending on the breach and the size of the company.
read also These 5 families are the richest in the world in 2023 AI Act, what are the prohibitions and obligations envisaged? Focusing more specifically on the AI Act, there are many prohibitions and obligations included within the regulatory system dedicated to artificial intelligence.
The list of bans includes: biometric categorization systems that use sensitive data such as political, religious beliefs and race; the untargeted collection of facial images from the Internet or CCTV footage to create facial recognition databases; emotion recognition in the workplace and in schools, social scoring and artificial intelligence used to manipulate or exploit user vulnerabilities.
The agreement then provides for a series of obligations for high-risk systems, including an assessment of the impact on fundamental rights.
AI systems used to influence the outcome of elections and voter behavior will also be included in this category.
AI Act, what are citizens' rights? In all this, in addition to providing sanctions for those who violate the law, obligations and prohibitions for companies, the AI Act protects the rights of European citizens, not only the right to privacy but provides that citizens have the right to lodge complaints about AI systems and to receive explanations about decisions based on high-risk AI systems that impact their rights.
Or this is why the Council representative during the negotiations, Carme Artigas, Spanish undersecretary for digital, said she was proud of the agreement, as it promotes innovation while protecting citizens' privacy.
read also South Korea grappling with the problem of porn espionage