Committee on Legal Affairs Ordinary Meeting - Extraordinary Committee meeting
Legend
Transcript markers
- text Sentence being played
- text Search result
- text Intervention cited by the link you opened
Click a sentence to play the video from that moment. Select text to quote it with the speaker, timestamp and link.
More options
JURI European Parliament
Follow its sittings
Sergey Lagodinsky
Here, we have to understand that this is not an isolated case of a single company. It is an issue of technology. And therefore, it is about the AI Act, and we think that it falls within the scope of Article 5, and we need a special prohibition. We can discuss this later or on a different occasion. Registration in the EU database for high-risk AI systems is not burdensome. The cost of this registration is under €100.
And at the same time, this proposal would mean that we will have much less. It's too early in the morning. Much less information about who, how do you say this? Transparency. Much less transparency about the companies that are excluded. There are many more proposals that we are discussing.
They are all in the papers. I don't want to take too much time or spam your brain in the morning with my proposals. I hope you read them carefully, and I hope that we get into a discussion either here or during the technical meeting with your teams. Overall, the timeline for the jury, as you know, is that February 12 is a presentation. February 13, until noon, is the deadline for amendments.
I'm doing the work of the secretariat now, I guess. And it's Strasbourg week, and then on February 24, there is the jury vote. So, it's a tight program.
It's all February.
Because we need to be ready before the lead committees start their work on substance, and we want to be ready with our position to give it to them. That's what I meant with the tight schedule. So, I'm looking forward to our discussion and our cooperation.
Ilhan Kyuchyuk
Good. Thank you very much. I would like to give the floor to Adrian Vasquez Lazara, who will be the usual suspect Axel was.
Axel Voss
So, thanks a lot, and good morning to all of you. So I'm reading out what Adrian has sent to me. So he's thanking the rapporteur for the work done, and he's considering this report a solid and constructive basis for our discussions. In his view, the Commission proposal achieves an appropriate balance between simplification and safeguarding the integrity of the AI Act. His priority will be to ensure coherence and legal certainty while avoiding the reopening of fundamental political compromises that were carefully negotiated. The amendments he will table follow the same approach, strengthening simplification without broadening the scope or disturbing the original balance of the act.
Adrian is fully supporting amendments 9, 24, 28, and 33, as they enhance clarity while remaining aligned with the objective of simplification. With regard to amendments one and twelve, he would urge caution against modifying the definitions to explicitly include AI agents. Such changes would amount to a substantial extension of the scope of the AI Act and could generate enforcement and implementation difficulties. The definitions were among the most sensitive elements of the original agreement, and I do not believe this. He does not believe this omnibus exercise is the appropriate forum to revisit them.
On sensitive data, Adrian favors remaining as close as possible to the Commission proposal. If we are serious about addressing bias and discrimination in AI systems, we must also ensure that developers have access to the data necessary to identify and correct these biases. He is supporting maintaining the commission's position on real-world testing, as well as the deletion of the register under Article 49. More broadly, I would preserve the core elements of the commission proposal. As for amendment 33, he believes it runs counter to the political agreement reached in the AI Act, particularly regarding differentiated compliance timelines for public authorities.
For that reason, he would be inclined to oppose it. Overall, he's supporting focused and proportionate simplification without extending the scope or reopening settled compromises. Thank you.
José Cepeda
Yes. I would like you to speak in Spanish, please. Bueno, Laura.
Thank you. First of all, I'd like to thank Sergei, the rapporteur, for all his work with some very interesting proposals. Now, it's true that from our point of view, Europe cannot dilute the regulations around AI, especially when things are becoming more radical technologically and geopolitically as well. The whole digital arena is very important. And this really goes against what Europe had developed at the time.
What we need to continue to do is safeguard integrity. The North American model is characterized by large companies with a more coercive approach than the European one. And then there's the technological authoritarianism that we have in China, their model, and how they define things. Europe needs to stick to this framework of rights and democracy. And it's essential to remember that artificial intelligence must not be dealt with as a product that can be regulated, but as a system that can affect employment, mental health, and health.
It's a multidimensional issue. And there's also the risk of breaking down the European model. Simplification is important, but not deregulation. Legal certainty and the protection of fundamental rights must not be barriers to innovation. It's clear that innovation needs clear rules.
Digital sovereignty is not based on exceptions, and we cannot see any backsliding. Therefore, this parliament should not delegate its powers for regulating technical issues. As the rapporteur was saying, there are some pernicious negative effects. Regulating deepfakes and protecting minors, women, and the most vulnerable people must be one of our priorities. And for this reason, I support the ban on the creation and use of sexualized images without consent.
We've seen a few recent examples with Grok. These can generate intimate images without people's consent, which is a serious breach of dignity and privacy, affecting specific sectors, particularly women and girls. In any aspect of our life, we cannot allow, for example, a child to drive a car or a car to travel without certain rules being in place. Artificial intelligence, therefore, must also have clear rules and regulations. And I think that this digital union is a great opportunity to adapt this to the reality of now.
AI cannot be a tool for abuse in the digital sphere. Europe must continue to protect our citizens' personal data and their fundamental rights as well. We will be tabling a series of amendments on this, and the report touched upon this. Generative AI, the move to AI without human control, needs to be restricted. One thing that worries me geopolitically, and we are working on this in all of the debates that we're engaged in, is security and defense.
AI is becoming a clear tool used against the protection shields in our countries. Cybersecurity is not mentioned once in the current regulations. This needs to be taken into account more. So, we are working on this. And of course, we are finalizing our amendments that we will table for this text.
I think that we have a broad space here where we'll be able to come to some consensus in order to ensure that artificial intelligence is about technological development, but at the same time, our citizens' rights need to be protected. Thank you.
Tobiasz Bocheński
Well, thank you very much. On behalf of ECR, we believe that the Commission's initiative is good enough, as it is another step towards simplification and deregulation. But of course, this step is not big enough. Like the previous speaker said, we need simplification. Let me add that we need deregulation.
There is one fundamental question that we need to answer. Parliament is talking a lot about the regulation of a great many areas. We are talking about AI-based applications like Grok or ChatGPT. So let me ask you this: why is there no European application? Why is there no European system?
Why? Well, due to overregulation, all European startups and all European people who are working on this are leaving Europe for the U.S. S. We are not competitive. We're not an island.
We are part of a global, interconnected network. Unless we draw conclusions from this and deregulate so that new technologies can be developed here in Europe, what will happen is that we will be left with regulations and rules but no influence, because there will be no modern AI models produced here in Europe. So, in general, we support the broad direction of the commission's initiative, but the commission's work is too shallow and too superficial. It has to be broadened and deepened if we want Europe to be competitive here. Of course, we can talk about individual rights, protection of property, intellectual rights, and all of this.
All these things are important, but this raises one very important question. What conclusions are you drawing from Draghi's report? Since we are not competitive, we have no competitive advantage here. When you go to any international forum which deals with the development of modern technologies like IGF, we, the European Union, are the laughing stock because we are deemed leaders in regulation but not in innovation. So, talking about such initiatives is very important indeed.
And we, ECR, believe that deep deregulation is needed because only freedom and liberty—academic and economic freedom—are the only things that can give Europe any sort of advantage, as we keep losing them day in, day out.
Ilhan Kyuchyuk
Mr. Buchenski, on behalf of the New Europe group, will be Madam Frank Lawrence.
Merci. Thank you.
Laurence Farreng
Thank you very much, Chairman. I'll be speaking French. First of all, thank you very much to the rapporteur for this very important work. And, of course, we've got very tight deadlines to work within. I must say, I don't fully agree with the objectives stated in this simplification text.
We've had a lot of trouble negotiating these files during the previous parliamentary term. We were the first continent to set a framework for the AI revolution. And I think the objectives we set at the time were absolutely correct. Of course, we were then hit by the arrival of generative AI, which completely changed the paradigm. So, I think that we should be proactive.
We should be financially supporting the EU AI sector, but I don't think this omnibus has found all the best solutions. For example, I don't think the stop-the-clock provisions give clarity to our companies because they reintroduce uncertainties and allow the commission to decide on the implementation date. As for the working method, I think that postponing by six months the labeling of AI-generated content is a serious error because, at the moment, we already have deepfakes damaging our democratic processes, and they are already very damaging for our cultural and creative interests. We should support copyright, and we should engage in that battle. Then, when it comes to properly managing the use of AI, too much is left to companies, whereas it should be up to governments to decide on the system.
And then, we also have the use of European data for AI training. That's a gift to foreign companies, particularly Chinese companies, rather than EU companies. So, like the rapporteur and the other shadow rapporteurs, I think we urgently need to respond to sexualized content that is targeted at women and young women in particular, without the consent of the users, by nudifiers. There are also other systems that attack the dignity of some of our youngest citizens. So, these are my priorities for this file.
I'd like to thank my colleagues in advance for all the work. I think we've got a good working environment on this file, and I hope for progress.
Ilhan Kyuchyuk
I see, Madam Farang. The list of shadow rapporteurs is complete. I would like to open the floor to any other member wishing to speak. Mister Deepavan?
Ilhan Kyuchyuk
It's not reflected in my notes because we are not notified. So, we have to make sure that we are notified.
Okay. Okay.
Ton Diepeveen
I'd like to say something in my own language. Yes, please.
Thank you very much, colleagues. And I'd also like to particularly thank the rapporteur, Mr. Lagodinsky. I'd really like to thank you for your work. As often, I'm surprised by the amount of energy that has been put into this file, although I don't agree with all the content.
I think that the AI framework needs to be simpler in the EU. There should be less bureaucracy, less red tape, and fewer limitations on companies that develop AI. We need to ensure that Europe remains attractive for investment. It's also important for our strategic autonomy, for employment, and for our international competitiveness. That's why we support the Commission's proposal entirely.
So, the commission can fully count on our support when it comes to simplification and omnibus measures. This also applies to the other omnibuses. Simplification really has to be the keyword. The rapporteur would like to go in a different direction. He's actually talking about new rules rather than simplification.
For example, for high-risk AI applications, the Commission is proposing simplification, six sixteen, and Annex 3 in particular. The rapporteur simply gets rid of that. And when it comes to CVs, well, that still counts as high-risk AI. The rapporteur reminds us that carers who lose their licenses for serious crimes, like sex crimes or serious abuse, can go on and work in another member state. Well, in cases like that, AI isn't a risk; rather, AI is a necessary instrument to try to prevent medical tragedies.
The same applies to our military. AI is going to be important for recruitment in terms of public security. Those are clear examples of where there needs to be a reworking of the rules, even for high-risk AI. I hope that the rapporteur will take account of these kinds of arguments, and I'll be entering a main amendment.
Ilhan Kyuchyuk
That is submitted to the secretariat. Okay. And then you will not be notified as the official shuttle on behalf of.
Ilhan Kyuchyuk
Okay. This will be solved. Okay. Okay. Okay.
I'm sorry. After the meeting, For ECR, Mr. Mantouvani?
Mario Mantovani
Yes. Thank you very much, Chairman. I'll just say a few words. I'd like to thank the rapporteur, first of all, for the digital omnibus. Certainly, it's an important step.
It reduces overlapping rules. It reduces bureaucracy, and it should benefit small and medium-sized companies. And it's particularly on the subject of SMEs that I wanted to say a few words. I think it's good that mid caps should be taken into account, but micro companies and startups should have particular protections that are proportionate to their requirements and their size. So, I just wanted to say to the reporter that for micro companies, there needs to be some kind of simplified system.
For example, exempting them from the most serious obligation. On high-risk applications, it could be useful to have a self-certification system. So the question is, if you have a sectoral certification, you shouldn't have to pay twice. So, it might be useful to move in that direction. Sustainable form of liability, particularly for small players, with a ceiling to limit it. And then, in many cases where you've got open source data, you shouldn't be liable for changes made by third parties.
Micro companies also need more time. Let's give them a little more time to adapt. Twenty-four months, for example, I think that would be a good penalty-free period that will allow companies time to adapt before penalties are actually applied. That could ensure that AI remains a source of value and innovation, even for the smallest companies.
Ilhan Kyuchyuk
Grazie. If there is no other request to speak on this matter, I would like to give the floor back to the rapporteur. Sergei?
Sergey Lagodinsky
Thank you very much. This was already a great start for our discussions. One thing I would like to say at the outset is that I think we all share the sentiment that this omnibus is something that should and would benefit our digital industry and our digital SMEs, and this is what we want. And at the same time, we want, of course, to strike the balance between that and the EPP. It was rightly said that the original will of the legislators, including Axel, myself, and many other colleagues, was reflected because we negotiated this law. However, I think that some of the groups will have to decide.
Do they want to be the legislators that legislate, or do they want to be the legislators that give only the commission the right to meddle on this law? And this is the problem. If you, on the one hand, Axel mentioned this. If you, on the one hand, support, for example, deleting or introducing real-world tests, something that was not agreed upon during our law. And on the other hand, you're saying that introducing other things, like, for example, the definition by us, is something that is not productive.
In this particular case, we have to decide: do we want the commission to be a legislator? The commission is not a legislator at all. We are those who can add things, and this is what we should do. But you cannot pick and choose what you want to choose and what you want to add. On the situation, what our ACR friends and colleagues mentioned is freedom and liberty.
And this is, of course, the cornerstone of the European Union, and this is what we are addressing. How can you be free? How can you talk about liberty? And if you can be undressed online without your free consent. And your picture as a porn star is being circulated everywhere without your free consent.
And we have to address this. And we have this unique opportunity to address this here and not wait until someone else starts dealing with this issue. So, from that perspective, but also from the perspective of dignity and morals, I think this is something that goes in the direction that our ECR colleagues and our Patriot colleagues probably share. So, from that perspective, let's think about these proposals from that point. Renew said very clearly, and we fully support this.
It is about more certainty for our small and medium-sized companies. And if you talk to them, they don't have problems with the AI Act. Really. Talk to me, I talk to scale-ups. I talk to startups.
They all say the AI Act is not our biggest problem. What they would probably have a problem with is if we say that we move and use the stop-the-clock, we move the deadline, and then we give the commission an opportunity to move it back and forth without knowing when the actual deadline is, and that's precisely our proposal. We are not big fans of stopping the clock. But if you introduce stop the clock, don't have an additional option for the commission to then revert from stop the clock and just do it at some other point. This is something that brings uncertainty.
So, colleagues, I think we all share concerns, and I hope very much that our teams, unfortunately, on the technical level, the negotiations will take place and will go in the right direction. I hope very much that we will address the issues. Some of those, we don't have an opportunity for, like liability issues. The commission has withdrawn the AI liability directive. This would have been the right place to discuss that.
But I'm very confident that we will find a good solution. Looking forward to our continued discussion.
Ilhan Kyuchyuk
Okay. Thank you very much. And the deadline for tabling amendments is February 13 at noon. The next part of the meeting will be in camera, so we need a couple of minutes of break.
Ask anything about this video — I analyse its transcript to answer.