Brando Benifei: „It's not a forced relocation''

This means they can become discriminatory, they can be biased, as we say, and also when we use them in contexts that are very sensitive, like administration of justice, like the development of children, like the decisions on whether a person has a right or not, for example, for a social entitlement. When you use AI in this context, you need to be sure that everything has been done to reduce the risks of errors, of malfunctioning, of damaging people. And this is why we have proposed an AI regulation. First in the world law, In fact, it's not inventing anything. It's putting together existing ideas around the risk mitigation that in other contexts are done, but not with the law.

They are voluntary, they are recommendations, so they are not very stable. The business, the public authorities, they already worked around this concept. But now we put it in a law. We want to become a universal approach towards AI that can make artificial intelligence compatible and acceptable in our democracy. We want people to trust that the European institutions and the public institutions in general, they have put enough protection for them in front of this huge technological change. Yes, but the technological change is ongoing. Does the AI Act prevent us from these challenges and risks in the future, when we don't even know where this goes?

Yes, I would say it's quite well prepared, because the AI Act is based on AI use cases, not on technology itself. This is important to underline. The approach of the regulation is to look ...at which use cases, so contexts where AI is used are more risky and regulate them accordingly. If they are less risky, less regulation. If they are more risky, more rules for the developers of these AI systems. This approach is flexible and also is adaptable to new technological advancements on AI. Also I need to say that the way the regulation is written, although it's not final today, we are still negotiating with the member states, it also entails the possibility of updating some parts of the law that can be in need of updating in the future,

With an easy procedure, not with the whole legislative work. So we have put some margins of flexibility and adaptation of this legislation for the future challenges that we might be in the situation to confront with. So we, the Europeans, can feel safe. I hope so. We are trying to do the best in this sense. I think we cannot completely eradicate any risks. If you think, even when you cross a street, there is a small risk that a vase from a balcony ends up on your head or There is a car running crazy that can hit you. But in this approach and this is the objective of this regulation we really try to make the AI capable of giving opportunities, of giving possibilities to make our lives better and reduce to the minimum the risks that it entails.

It's the EU regulation while AI knows no borders and so the dangers it might entail don't. Is there international cooperation that addresses these omnipresent risks? Probably as for example most European vehicles are made in China and this country is focusing on technological development very successfully, we should have them on board in such a global compact for AI. It's crucial that we proceed with global cooperation. In fact, we are discussing today in the European Parliament, in this moment, this issue of the global cooperation, which is extremely relevant. It is also the other aspect.

We need to regulate internally, but we also need international cooperation. This is especially concentrating today on one aspect of AI, that is the generative AI. CzaGPT, Midjourney and such, which are particularly innovative and also pervasive because people are using them more and more and so they have a big impact in our daily lives of people, on society, on the way work evolves. So we need common rules, we need a minimum, but also at the same time we want to have our own legislation because you need to consider that if you are American, Canadian, Brazilian, European, Chinese.

You will follow the same rules if you want to enter the European market with your AI system. So it's not on one end penalizing Names mentioned''. Names mentioned''. Names mentioned''. AI scientists and other notable figures issued a one-sentence statement, mitigating the risk of the extinction from AI should be a global priority alongside other societal scale risks such as pandemics and nuclear war.

Sounds really scary. Do you agree we should fear the progress of the AI? I don't think so, because this progress will not be stopped, so we need instead to govern it. On one hand, I totally understand and respect the attention on the systemic risks, on the big global risks of AI, which means that Global cooperation, even a global treaty on AI needs to be done. Like we did on pandemics, we did on the commerce of weapons. It's important that we do this, but this should not distract us from the daily risks of AI for the common people, which is another dimension. More specific, smaller to some extent, less relevant globally but it's about the lives of people that might be ruined by a bad AI but instead they can be better.

These lives can be better if AI is used in the right way. So we also need to look at this domestic and daily governance of AI that is the objective of our regulation. Karolina Zbytniewska, European Media and Information Fund, EURACTIV, Poland, European Union. I think we should have the right approach looking at this topic. European Union is a union of countries that has a dimension of solidarity, which was demonstrated in these years by the cohesion funds, for example, that were extremely important for the new countries that joined in the last 20 years.