Last April, the European Parliament passed the EU Artificial Intelligence Act, setting a global precedent for regulating this evolving and complex field.
This new legislation aims to protect fundamental rights, democracy, the rule of law, and environmental sustainability from high-risk AI systems, while simultaneously fostering innovation and showcasing Europe as a pioneer in the sector.
Romanian MEP and Renew Europe group member, DragoÈ™ Tudorache, one of the regulation’s rapporteurs, spoke to Phileleftheros emphasising that this is a pioneering regulatory framework that protects Europeans from the risks inherent in technology. “This technology has a transformative impact on us all—on society, the economy, and geopolitics—and we cannot afford to ignore its development and interaction with civil society,” he noted.
Once ratified by the European Council in the coming months, the regulation will become EU law, with full implementation expected approximately 24 months after it takes effect. It sets different obligations for AI systems based on the potential risks and implications of their use. “Most importantly, we did not attempt to regulate the technology itself, as that would be impossible. We examined its uses to identify areas where its operations pose risks to us and society,” the Romanian MEP explained, noting that the new rules prohibit certain AI applications that threaten citizens’ rights and contravene EU principles.
Why do we need rules to regulate Artificial Intelligence?
Regulations are necessary because AI is a transformative technology impacting society, the economy, and geopolitics. We cannot overlook its development and interaction with civil society. Although AI holds tremendous potential and promises significant benefits, it also presents risks—for example, when obtaining a bank loan or when schools use algorithms to place children in classes. The question is, can we trust those developing AI to mitigate these risks adequately? History, such as with social media, has shown that companies were neither quick nor willing to address systemic risks in technology’s interaction with society. Hence, it was deemed necessary to adopt strict rules to ensure the smooth operation of Artificial Intelligence.
What are the key provisions of the Regulation adopted by the European Parliament last April?
The most crucial aspect is what we call a risk-based approach to Artificial Intelligence, which essentially means we didn’t try to regulate the technology itself—that would be impossible. We examined its uses to pinpoint areas where its operations pose risks to us and society. The regulation leaves the vast majority of AI that optimizes industrial processes, traffic, transport, and air traffic management, which do not endanger our rights, unchanged. Regulation is present where there are risks. At the core of the Regulation is the categorization of risk levels. Systems posing unacceptable risks are banned; high-risk systems will be subject to strict rules, while those posing lower risks will have more lenient regulations.
When will the Regulation start to be implemented, and what do Member States need to do to comply?
We expect the Regulation to be published in early June. Subsequently, it will begin to be implemented in three phases and should be applied within two years from now. Thus, Member States will have 24 months to align and establish the relevant Authority, either by modifying what they already have or by appointing a new one responsible for implementing the regulation.
Do you believe the EU’s pioneering effort to regulate AI will make a difference even if other countries do not adopt similar rules?
Currently, we are the only ones who have set such rules. However, I am convinced others will quickly follow. I’ve interacted with several national parliaments and governments around the world—in Latin America, Asia, and other regions—and learned that they are at various stages of preparing their national legislation on AI. Here, I think we have an advantage because our rules will serve as inspiration for other countries. Within two years, mandatory standards will exist, and developments in AI, both within and outside the EU, will be influenced by these European standards.
Technology advances very rapidly. Are you concerned that by the time the regulation is applied, it might not be as relevant?
Not at all because we have the capability, through a protocol, to regulate technical issues very quickly as needed. Thus, with the new regulation, where changes are required, they will be made continuously. This way, the new regulation will be able to adapt continuously and keep pace with technology.
Do you believe that technology companies, given their power, will try to stop or weaken the AI Regulation?
Well, they already tried to do something similar when discussions started. For a long time, they hoped we would not succeed. Now that the legislation has been adopted, they realize there’s no way to avoid it. Whether they like it or not, they understand that this is the new standard for AI and they must adapt to it. Will they raise objections? It’s natural for that to happen. The question is whether the legislation will allow them, at a European level, to bypass the rules. I don’t believe that will happen because there will be very strict penalties.
Could the regulation act as a brake and hold back the EU from technological development, especially in competition with China and the USA?
Absolutely not. I don’t believe there’s a dilemma between regulation on the one hand and innovation on the other. Rules haven’t stopped pharmaceutical companies, automotive industries, and all other sectors from developing. Rules do not inhibit innovation, and a Wild West regime doesn’t particularly help. Therefore, I believe the existence of standards actually strengthens the companies themselves as it creates market stability and allows them to operate in good faith. And this is the feedback I receive from technology companies.
