Finding a balance between protecting fundamental rights and fostering innovation, drafting laws capable of keeping pace with the artificial intelligence (AI) innovation. These are two of the challenges encountered by MEPs Dragoş Tudorache and Brando Benifei, the co-leaders of the EU AI Act.
ADVERTISEMENTOn March 13th, the European Parliament approved the AI Act with an overwhelming majority: 523 votes in favour, 46 against, and 49 abstentions.
The regulation, expected to enter into force in May 2024, is already a historic landmark. Despite generative AI already being part of our lives and making headlines in international media, the European Union (EU) is the first-ever institution to approve a set of laws to regulate AI systems, while protecting people’s fundamental rights.
To better understand this ground-breaking regulation, this special episode of Euronews Tech Talks brings you the experience of the two lawmakers co-leading the AI dossier in Strasbourg’s Parliament: Dragoş Tudorache and Brando Benifei. The former used to work as the Interior Minister of Romania and represents the liberal Renews Europe Group. The latter sits with the socialists and was one of the youngest Italian MEPs ever elected.
How did they work together? What challenges did they face?
The genesis of the EU AI Act
The story of the EU AI Act starts way earlier than the launch of the popular tool ChatGPT in 2022. “The European Commission had a high-level expert group on AI, its applications, and potential risks before 2019,” Tudorache explains. “Then, in 2019, President von der Leyen announced there would have been a legislative proposal on AI. In 2020, there was the launch of a special committee and in 2021 we started to unpack the concept of AI,” he adds.
Since the very first step of the negotiation, the main challenge was clear: make sure that the fast and unpredictable development of AI technologies would not outdo the regulations. To face this issue, the lawmakers came up with a strategy: “We had to make sure that the obligations were technology-neutral: If you define obligations of transparency, those will stay relevant, no matter how complex the algorithms will become,” the liberal MEP underlined. Together with this methodology, also a juridical technique: “Whatever new uses would arise with new risks that might occur, they can always be added to annex three, which is the list of high-risk applications of artificial intelligence, which we have deliberately left open,” Tudorache adds.
More precisely, the AI act follows a risk-based approach: the riskier the AI application, the more scrutiny it undergoes.
“We have limited the use of real-time biometric cameras in public spaces to the pursuit of suspects of very significant crimes, but we have also banned emotional recognition in workplaces and schools,” the Italian MEP adds.
But how can the protection of human rights not impede technological development?
“The purpose of this legislation is not to stifle innovation, but rather to build trust. Our model, where consumers are heavily protected and human rights are central, will not be hindered by this disruptive technology. Instead, it will be integrated to develop our AI model,” says Brando.
