Green light from the Eurochamber to the AiAct, the text that contains the new EU rules for artificial intelligence

The Strasbourg plenary approved with 499 votes in favor, 28 against and 93 abstentions, the first regulation in the world designed to give technologies such as Chat Gpt respect for EU laws and fundamental values ​​and to combat digital discrimination and prevent disinformation and the use of «deep fakes».
The vote paves the way for trilogues with the member countries and the EU Commission, the first negotiations are already scheduled for this evening. The European Parliament then supported the total ban on the use of biometric recognition technologies with artificial intelligence in real time in public places. The EPP amendment which asked for some exceptions to be included in the AiAct to allow the use of facial recognition cameras on the streets and at EU borders for reasons of national security or in cases of missing minors was rejected. The rules follow a risk-based approach and set obligations for AI system providers and operators according to the level of risk they may generate. AI systems that present an unacceptable level of risk to people’s safety, such as those used for social scoring (classifying people according to their social behavior or personal characteristics), will therefore be banned. The text approved by the European Parliament provides that the classification of high-risk applications also includes AI systems that cause significant damage to the health, safety, fundamental rights of people or the environment. Artificial intelligence systems used to influence voters and the outcome of elections, and recommendation systems used by social media platforms (with over 45 million users) have been added to the high-risk list.
Generative AI systems based on such models, such as ChatGPT, according to the text, should comply with transparency requirements (declaring that the content has been generated by AI), also helping to distinguish so-called deep-fake images and from real ones, and provide safeguards to prevent the generation of illegal content. Detailed summaries of the copyrighted data used for training should also be published. “We have seriously discussed the issue of using artificial intelligence technologies for border management. Our decision was that this regulation must also be applied in this area and this means obligations of transparency, and impact assessment on human rights”. Thus, Dragos Tudorache (Renew), who together with Brando Benifei of the Democratic Party was the rapporteur of the text on AI in the European Parliament. “When it comes to migration, we believe that even there there is a limit to the use of AI, we are in an area of ​​great risk and this must be taken very seriously, we must remain vigilant to ensure respect for fundamental rights”. added Benifei speaking at a press conference in Strasbourg on the green light to the new rules for Artificial Intelligence. “We have managed to keep a clear ban on facial recognition in real time – explains Benifei, head of delegation of the Democratic Party. There had been an attempt to politicize it, to turn it into a propaganda tool, but we won in Parliament and managed to maintain a clear safeguard to avoid any risk of mass surveillance.” “At the same time – concludes Benifei – we have maintained the possibility of non-real-time facial recognition to prosecute criminals and face risks that we may have in society”.