Committee on Constitutional Affairs Ordinary meeting
Legenda
Marcadores da transcrição
- texto Frase em reprodução
- texto Resultado da pesquisa
- texto Intervenção citada pela ligação que abriu
Clique numa frase para reproduzir o vídeo a partir desse momento. Selecione texto para o citar com o orador, a marca temporal e a ligação.
Mais opções
AFCO European Parliament
Seguir as suas sessõesEmmanouil Kefalogiannis
The investment gap with the United States is becoming too large to ignore. Between 2021 and 2027, business investment in the United States is expressed to increase by around 40% in real terms.
In the euro area, the increase is only 12%. This gap is increasingly driven by artificial intelligence. In 2026 alone, Google, Meta, Microsoft, and Amazon are expected to invest more than 700 billion euros, much of it linked to the artificial intelligence infrastructure. Europe cannot regulate technologies that it does not develop. This is why European AI gigafactories and investment in computing capacity are strategic projects.
They concern our ability to remain a global actor in the age of artificial intelligence, to strengthen European competitiveness and capacity, to develop our own AI. This need is clearly reflected in our report. Second, we need international cooperation. This is why our report raised the need to explore a global framework for artificial intelligence governance, inspired by the experience of international cooperation in the nuclear field. Artificial intelligence has no borders, therefore Europe should lead efforts towards an international convention for artificial intelligence based on common principles: transparency, safety, and human control.
Third, we must protect our democracy and our elections. This is another central element of our report. AI is amplified disinformation, deepfakes, foreign interference and manipulation. In our report we avoid presenting AI only as a threat. We need balance.
That means protecting democracy while using technology to defend democracy. Fourth, Parliament itself must have the capacity to follow these developments. This is why our report maintains the proposal to consider establishment of a parliamentary AI observatory while ensuring it does not duplicate existing European structure. Parliament needs its own institutional capacity to understand how artificial intelligence affects democracy, decision-making and European integration. Dear colleagues, at the same time we should never forget the social dimension, which is also an important part of our report.
Artificial intelligence must serve people, not the other way around. AI should support and empower workers. It cannot and must not replace the human beings. Human responsibility and human judgment must remain at the center of our workplace. Transformation of our economies must be accompanied by investment in skills and education.
Technological progress must strengthen European society rather than create new inequalities. Dear colleagues, our objective throughout this negotiation has been clear: a Europe that protects without being afraid to innovate, a Europe that regulates but also invests, a Europe that does not simply consume artificial intelligence developed elsewhere, but has the capacity to develop its own artificial intelligence.
As conclusion, these priorities are at the heart of our report. I believe the compromises before us move in the right direction. And I thank all shadow rapporteurs once again for their constructive contribution and cooperation. Thank you very much for your attention.
Sven Simon
Thank you very much, Mr. Rapporteur, and I now give the floor to the shadow rapporteurs for three minutes each, starting with the shadow rapporteur from the Socialists and Democrats, Vytenis Povilas Andriukaitis.
Vytenis Povilas Andriukaitis
Thank you. Thank you very much. First of all, I want to thank the rapporteur, Emmanouil Kefalogiannis, dear Emmanouil, you did a good job, and fellow shadows for good collaboration, as well as the secretariat and our teams for a great work in this important file. The secretariat and teams shows really very cooperative spirit. The rise of artificial intelligence is a force reshaping how the EU institutions function.
It raises real questions about where human judgment ends and automated interference begins in processes that carry constitutional weight and derive legitimacy. And this internal dimensions needs to be addressed.
Without human control, no constitutional ground. This presents genuine opportunity: faster and more consistent legislative scrutiny and drafting, more accessible multilingual cooperation. But also presents real threats to institutional accountability, the balance of powers between institutions and Member States, and to democratic legitimacy. These consequences touch the allocation of competences, the transparency obligations of EU bodies, and the accountability chains that underpin the Union's democratic character. They require rules on the use of AI, on human oversight, so that the EU's internal governance keeps pace with and sets a credible example for the external regulatory standards it is asking to the rest of the world to adopt.
I am happy to our report is that our report is a step toward the right direction.
We made sure AI use in the institutions stays strictly human-centered. Human oversight of AI tools, a right to explanation and human review of automated decisions will require regular parliamentary AI observatory reporting to AFCO, specifically on institutional constitutional implications, keeping Parliament, not external providers, as the ultimate actor. We also made sure to go towards stronger, coordinated European institutional capacity to prevent fragmentation and dependency by reinforcing Parliament's scrutiny power over the EU's global AI engagement by enhanced access to information and regular reporting and call for any inter-institutional coordination mechanism to be built on transparency and timely involvement of Parliament, rather than informal agreements. This is paired with support for building genuine EU-level analytical capacity, so the institutions aren't reliant on uneven national capabilities or external foreign-designed systems, addressing both inter-institutional fragmentation and the risk of unequal AI capacity between Member States.
We also secured equality alongside human dignity as an explicit value that the AI governance must uphold, backed budgetary authority safeguards for Parliament, and pushed transparency and privacy commitments as binding conditions on the EU's external AI positions. The broader package this feeds into also insists responsibility for AI system effects must not be unfairly shifted onto end users, consumers, or workers, so that AI economics gains do not come at the cost of labor market fairness or unequal protection across the Union. Thank you.
Sven Simon
Thank you very much. The shadow rapporteur for the Patriots for Europe, Ernő Schaller-Baross.
Ernő Schaller-Baross
Dear Chair, dear rapporteur, dear colleagues. First of all, I would like to thank the rapporteur for his work on his report and in particular for his continuous openness to dialogue and compromise. As highlighted earlier, I believe that technological developments, including artificial intelligence, must be addressed in a pragmatic and legally coherent manner. In this context, I particularly welcome the deletion of the reference to a possible treaty change. I would also like to thank the rapporteur for taking this point into consideration and for deleting his previous proposal in this regard.
In our view, this represents the best possible compromise.
My group's priority is the proper implementation and enforcement of the legislation we have already adopted, rather than reopening fundamental institutional questions. Let me also address the proposal for an additional AI observatory within the European Parliament. While we recognize the rationale behind this proposal, Patriots remain cautious about whether a new institutional structure is necessary at this stage. My group is not yet fully convinced that an additional parliamentary body would provide sufficient added value, particularly given the structures and the mechanisms that are already in place. There is also a risk of creating unnecessary bureaucratic layers or duplicating existing responsibilities.
Overall, I believe that the report is moving in the right direction and that amendments have contributed to improving its balance and focus. A few elements should nevertheless be approached with caution, ensuring that we do not go beyond what is necessary and proportionate to address the challenges at hand. Finally, I would like to thank the rapporteur once again for his work. I believe we had a thorough and detailed discussion of this topic and we have had the opportunity to carefully consider the different perspectives and concerns of every parliamentary group.
I would also like to thank the IMCO and LIBE committees for their contributions and work on this report. We look forward to the committee vote and to the next steps in this process. Thank you very much.
Sven Simon
Thank you very much. Now for the European Conservatives and Reformists, Adrian Ashinia is replacing George Piperea, who was the shadow rapporteur.
Adrian-George Axinia
Thank you. Since Mr. Piperea couldn't join this meeting today, I was kindly asked to say a few words on his behalf. I would like to start by thanking the rapporteur, Mr. Manolis Kefalogiannis, and his team for his great work on this report and the very fruitful collaboration on this. As you probably know by now, Mr. Piperea was working on this file with a dual role, that of lead rapporteur of the IMCO opinion and that of shadow rapporteur for the ECR on the AFCO text. The shadow rapporteur believes that we have a good text as it stands today, and he very much welcomes the input by both committees, LIBE and IMCO.
Of course, having worked on the IMCO opinion himself, he highly appreciates the fact that the compromise text has taken on board many of the matters reflected in the opinion.
Needless to say, walking the line between progress and protecting our citizens is no easy task and we have to proceed wisely with both caution and ambition. And this is exactly what this report has done. And we are glad that it starts as an AFCO report as we should treat AI from a constitutional viewpoint first and foremost and expand further along the way. This is why IMCO and LIBE have a great role in fine-tuning the text, even though the ramifications of AI touch upon each and every aspect of our citizens' lives nowadays. More concretely, we welcome among many other aspects the fact that it seems to be agreed across the political spectrum that we should prioritize human scrutiny.
No AI system should be able to replace human oversight and we all agree that humans must have the final say in AI-related processes. Our shadow rapporteur also welcomes the dedicated space for protecting human rights in this report. As we have seen, AI can facilitate human rights violations in various ways, ranging from algorithmic manipulation and discrimination to the creation and dissemination of deepfakes and the theft of personal data or even one's visual identity.
Furthermore, we are glad to have a text that also includes provisions relating to the effects of AI on electoral systems. This is the key for us as a directly elected body. We should watch that our citizens' political opinions do not fall prey to the ruthlessness of algorithms and AI manipulated content and that all these technologies are not exploited in order to interfere with the integrity of electoral processes and undermine trust in our democratic institutions. Lastly, mechanisms of redress and accountability are highly important and we are glad that the text remains these. Without them, our citizens' rights risk being rendered meaningless.
Obviously, there is much more to be done in terms of AI regulation, but from AFCO point of view, we think we can achieve already a substantial result by voting this report with a large majority. Thank you.
Sven Simon
That sounds quite positive. So for Renew, I guess Sandro Gozi is replacing Helmut Scholz.
Sandro Gozi
Yeah, thank you, Chair, and warm thanks to the rapporteur who's done an excellent job.
We are very pleased to have put on the agenda the issue of artificial intelligence from a governance and constitutional perspective. We talk about something, a revolutionary activities, opportunities, event, which is at the heart of our societies and therefore should be at the heart of the preoccupation, of the attention of the European institutions. And our reasoning, our contribution to the work of the rapporteur and also our amendment goes around three fundamental principles, which is rethink, protect, and promote. We have to rethink the way our institution work around the issue, the challenge, the opportunity of artificial intelligence, and this is why I think the report brings an added value to what we've done so far in the European Parliament from a broader, horizontal, constitutional perspective. We have to protect because artificial intelligence is a challenge and sometimes a threat to fundamental principles and to some of our priorities as European Parliament and Constitutional Affairs Committee.
I think of public security, I think of democratic accountability, I think to the protection of the democratic processes, to the constant disinformation and foreign interference, which aims to destabilize our institutional democratic structure and which is mainly done, not only, but mainly done through artificial intelligence.
So I mean there is a lot of work to do to rethink and to protect the way our democracies and our institutions, especially our European institutions, work. And the third point is to promote. I thank again the committee for the mission to the United Nations we had in July. It was a mission aimed to focus also exactly to the work that at global level, at United Nations level, they are doing around artificial intelligence. I think that all the participants to this mission fully grasped the importance of a European role on the international scene.
We are seeing a new competition around artificial intelligence at global level between a European model, an American model and a Chinese model. And I think that the promotion of the European model should go also through an institutional, a new proposal at institutional and governance level at global level. There is a work which has been done at the United Nations on the UN Global Digital Compact.
I think that as European we should take an initiative to strengthen international cooperation, to put flesh of the bones of the current global governance. And this is why we warmly thank, I warmly thank the rapporteur for having taken on board our amendment which aims also in this context to propose as Europeans, as European Union, at international level or at UN level the adoption of an international treaty on non-proliferation of high-risk and harmful artificial intelligence systems with a view to preventing the uncontrolled development and deployment of particular dangerous AI capabilities and ensure effective global safeguard and oversight. Rightly the rapporteur has underlined this. We have to learn from the approach we have used on nuclear issues and we need also to think on non-proliferation. AI provide us with huge opportunities but also with huge risks.
And this huge risk should be tackled through a non-proliferation approach at international level. So I hope that this proposal will be follow up and we should insist to follow up because we would like as Renew to see a European Union initiative at global level to promote this international treaty. So all in all I really, we really hope that there will be a large support. I'm pleased to hear that so far all the groups are supportive because this report deserves a strong support but also deserves the same commitment to ensure a proper and effective follow-up. Thank you, Chair.
Sven Simon
Thank you very much. Is there any member of the Greens who wants to take the floor? No.
And then a member of the Left? Oh, sorry. Excuse me. Colleague Orlando, please.
Leoluca Orlando
Thanks, Mr. Chair, we speak Italian. Desidero dire grazie al relatore Kyrtzos per l'ottimo lavoro e agli altri shadow che hanno contribuito sostanzialmente a definire una linea che ha un punto di partenza che costituisce la sfida ma anche il limite di questo lavoro. Lo accennava il relatore, noi siamo in presenza di una sovranità europea che non c'è e siamo in presenza di una sovranità regolamentare che non può esserci senza una sovranità tecnologica. E quindi l'assenza di una sovranità tecnologica rende sostanzialmente impossibile una sovranità regolamentare. Dai data center sino alle piattaforme sono condizioni perché si possa avere anche un potere regolamentare.
È questo il senso delle indicazioni che vengono date in questa relazione nella quale si fa riferimento all'esigenza di una particolare attenzione con riferimento al funzionamento dei sistemi istituzionali che è poi anche il funzionamento della democrazia, il funzionamento delle elezioni. E questo sicuramente costituisce una sfida perché in questo momento c'è una debolezza cronica dell'Unione Europea sul versante tecnologico in materia e quindi questo rende sostanzialmente l'Unione Europea debole rispetto agli altri competitor che sono oggi competitor, domani possono essere avversari o nemici a seconda del diverso svolgersi delle condizioni. Allora ecco credo che sia importante far questo, andare avanti su questo lavoro sapendo che non possiamo rassegnarci all'idea che venga soppresso ogni ruolo dell'intelligenza umana. Perché l'intelligenza umana a differenza dell'intelligenza artificiale ha dubbi, mentre l'intelligenza artificiale non ha dubbi. L'intelligenza umana per fortuna commette errori, mentre l'intelligenza artificiale purtroppo non commette errori.
E l'intelligenza umana presuppone responsabilità, mentre invece l'intelligenza artificiale non conosce il principio di responsabilità. Credo che basterebbe questo per dire come sia assolutamente necessario garantire un controllo umano esposto agli errori per difenderci da un'intelligenza artificiale che errori non ne commette e responsabilità non ne ha. Basterebbe citare le armi automatiche letali del progetto Agile per rendersi conto di quali siano i rischi connessi ad affidamento ad uno strumento che non commette errori la vita e i diritti delle persone. Grazie.
Sven Simon
Thank you very much. Now the shadow rapporteur for the ESN, Christine Anderson.
Christine Anderson
Yeah, thank you, and thank you rapporteur for your work on this file. Regrettably, none of our fundamental concerns about this report have found their way into the compromises. So let me just restate what I just believe to be obvious.
Artificial intelligence may be a new technology, but it certainly should not constitute a new competence for the European Union. Yet throughout this report, technological change becomes an argument for institutional expansion: a new observatory, monitoring for elections and political parties, a global AI treaty, even the suggestion that AI demonstrates the inadequacy of the existing treaties. This reverses the constitutional logic of the European Union. The treaties define what the Union may do. But the arrival of a new technology does not expand those powers.
And then you want to create an AI observatory inside the European Parliament and then give it a role in monitoring political parties and elections. What I see, it could actually pose a real problem for democracy. Of course we should address fraud, foreign interference and genuinely deceptive uses of AI. But who decides when political communication becomes disinformation or manipulation? Political speech, public debate and elections are not things for an EU institution to monitor simply because AI is involved.
That is the reason our amendments proposed a deliberately different approach.
Use AI to strengthen Parliament's scrutiny and to analyze the implementation of existing EU law, but never replace political or legal responsibility. Keep any AI observatory focused on supporting Parliament's existing functions, not monitoring political parties, elections or public debate. And address technological change within the competencies conferred by the treaties rather than using it as an argument for new powers or new treaties. AI should be a tool for democratic institutions. It should not become a tool for expanding its powers.
Thank you.
Sven Simon
Thank you very much. This closed the round of shadow rapporteurs and now the next speaker will be Klára Dobrev from the S&D group.
Klára Dobrev
Thank you. Thank you very much, Chair. Well, let's first admit to all the rapporteurs and shadow rapporteurs that I don't think that ever in human history legislators were in such a difficult situation as we are now because we have to regulate something what we do not understand. But it's not only us. If you listen to the tech gurus, the fathers of AI, they do not understand fully what's happening and their prediction is absolutely useless not for the next ten years but for the next half a year or one year or two year, what will happen.
So that's the reason why people all over the world are raising the question that is AI a blessing or it will destroy us? And this question is raised day by day by a lot of people. Here in the AFCO committee, we can discuss and debate about laws and institution protecting the European democracy, but let me, please let me tell you another aspect. Because these debates are going to be shadowboxing if we are not able to change the way how we approach AI, fundamentally change the way how we are approaching AI. The political right and liberals are using the term AI and the whole problem as it was a problem of the market, of GDP, big corporation, profits, as if it was something like a magic wand in order to increase European competitiveness.
And they deal with it like a market issue, like free market issue. To tell you the truth, I see absolutely the opposite way. We are far beyond the point where we could let the market to decide about the future of AI, far beyond that point.
AI has changed the way how we live, how we work, how we think, how we travel, how we eat, within a few months, fundamentally changed it. Algorithms decide even how we see the world, and of course manipulation, political manipulation became an everyday experience in our lives. So public opinion is constantly being manipulated by algorithms, sometimes not even people. And very soon we are going to see millions of people losing their jobs, they are losing it already, mainly entry-level jobs, young people, unemployment. And on the other hand, you will see fortunes by the big corporation you have never seen in the human history before, and not only fortunes, but political power, political power which can be used for blackmailing democratic institutions, and they are doing it already.
So that's why I'm afraid that this is going to be something like the end of the social peace. And if social peace disappears, then democracy itself will no longer be possible.
So tell me what more we can think about the institutional implication of AI if it's threatening the core democratic system. So my dear colleagues, if you ask me, then our first step must be to take the decision about future of AI out of the hands of the big corporation. This is the first step, and then we can discuss institutions and regulations. And to all nationalists, this cannot be done on a national level, definitely. This cannot be done on European level, definitely.
But we as European can be at least the frontrunners in order to start thinking about it. And to tell you the truth, we can only pray that it's not too late anymore. Thank you very much.
Sven Simon
Thank you. Is there any other speaker who would like to take the floor? If that is not the case, then I must say this report has a real chance of being adopted with a very broad majority. I think rarely has a rapporteur received such widespread praise, so congratulations already. And you have the final word if you want to comment or elaborate a little bit on what colleagues said, or you don't have to.
Emmanouil Kefalogiannis
The only thing I would like to say is to thanks once more all colleagues, and especially shadow rapporteurs for their constructive ideas we tried to include in our compromise amendment. And I would like to say a big thank to the advisors of the political parties, and of course the secretariat and my team, because they work for months to find a common ground. And I take very seriously into account all the worries that our colleagues raised in this discussion. If we can include last moment something, we are still open to discuss some radical ideas, either from the parties or from our colleagues. It's really a very important issue for all of us and of course the European Parliament.
We believe in Europe, we believe in what we can do in this institution, so let's go ahead and find common ground to get the best out of this proposal. Thank you very much.
Sven Simon
So thank you very much. I announce that the vote in committee is foreseen to take place on 1st October, which is our next committee meeting. Any other business? If that is not the case, it was a short committee meeting. Next meeting will be on the 1st of October, 9 o'clock.
The meeting is closed.
Otherwise let's pray.
Faça uma pergunta sobre este vídeo — analiso a transcrição para lhe responder.