Special committee on the European Democracy Shield Ordinary 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
Offset
Transcript offset
Export
Replay
Versions of this video
Transcripts may differ slightly from one version to another.
EUDS European Parliament
Follow its sittingsNeville Bolt
Thank you. Please allow me to thank the chairs and also everybody in Parliament for inviting us today. It's a great opportunity to spend time with you and also not just to share with you our ideas but actually to learn from you what your ideas and proposals are. And you mentioned the Next Generation Information Environment report that was presented to the Munich Security Conference in February of this year. And this is it.
And we can provide this to anybody who wishes to read it in a quiet moment. But I think what we should do is just take a step back from this report and try to look at the, first of all, the background and the context. So we convened for the whole of last year a number of experts from around the world. And they came from every possible field and discipline you can imagine. There were economists, there were historians, there were philosophers, there were sociologists, there were psychologists.
They came from banking, they came from government policymaking. And we did that very deliberately because ordinarily what you tend to get in expert research projects is a lot of similar-minded people coming together. Here we had people who didn't know each other, didn't share even the same conceptual language, which I think is really important because one of the problems that we find in political debate is how many people are talking past each other. They're not actually speaking to the same topic. This is especially important when we're talking about emerging technologies because this is a, it's a kind of umbrella concept.
It's enormous, particularly in its makeup but also in its consequences and implications.
So we first I think it's important if we kind of go through three different kind of stages, levels. One, maybe what I can do is offer a couple of thoughts on the political economy of the next generation information environment, which actually is almost another way of saying what is the future of liberal democracy? And then we'll go to what happens inside the technologies, and then we'll go even further into the operationalizing of technologies in particular contexts. So for us the overarching context was this. We're living in a world, in a kind of vortex of securitization and militarization, number one.
Number two, we're living in a Europe of low economic growth, which inevitably leads to popular grievances, dissatisfaction, and a feeling of disenfranchisement in major parts of populations. Three, we're seeing a we're witnessing an historic fracturing of the liberal democratic consensus. Four, we're seeing an acceleration in different kinds of technologies at a particular historic moment. And then fifth, we're seeing a direct attack on international lawmaking and the recognition of international norms. So it was interesting listening to so many different voices from so many different disciplines who were trying to address the same question that you're looking at today.
Some people said the problem is this area is full of tensions. Others said let's be more specific, let's call it contradictions. Either way, you're looking at a lot of particular dilemmas at a governmental level, whether national or international. I'll just list two or three of them before I hand over to Alina. One is this idea that we are the West is caught in an arms race, a kind of artificial intelligence arms race which is essentially between two parties, one the United States, the other China.
And Europe is talking about strategic autonomy or technological sovereignty, similar thing. I can tell you what our bankers told us, and I'm talking about, you know, the biggest banks in the world like Goldman Sachs. I can tell you that they said if you believe that Europe is still capable of building its own European AI stack, forget it. The train has left town, you're too late. All you can do is in the stack where there are different technologies that contribute to an integrated vertical system, Europe can continue to use its expertise, its excellence in research and development and fit into that.
The problem with that is of course that Europe risks becoming technologically colonized, either by the United States, probably United States, or China. So that was one of the big dilemmas that was identified. Another was the fundamental difficulty of that faces the European Union in one respect. What was identified as overregulation, overlegislation, overbureaucratization, albeit to protect the choices and the rights of individual citizens, nevertheless in the face of an ever-accelerating set of technologies which the consumer marketplace, not governmental gatekeepers, but the consumers want.
They're making the choice far, far faster than government representatives. A third, a third idea that many people focused on was, just suppose Europe wanted to compete against China and the United States in shaping the information environment through owning its own ethically rooted artificial intelligence stack, what's the problem? Fundamental problem is that European financial capital markets are too risk-averse. So the dilemma is, how do you incentivize private investment in public-private partnerships on such a vast scale? How do you incentivize them through what kind of return on their money?
How do you allow them to make profit? Because at the moment, that's what they can do in the United States. The money goes to the US capital markets. We're all aware that whatever it is, it doesn't really matter, but up to 60% of Wall Street is invested in these emerging technologies. So how do you compete with US markets, capital markets?
Another point that came across was, just suppose you could, just suppose you could, and you invest in a number of very promising startups, you build them, you give them the regulatory space, the fiscal space in which to operate. What is to stop in an open marketplace American companies, Chinese companies, possibly, but let's assume that they can be kept out, but American companies just coming along a year or two later and buying them up, which is what happens at the moment anyway. And a final point before I hand over, which is the fear of being left behind. And this is a big fear that we all share. And yet we're talking about how do we compete in this notional arms race for emerging technologies, while at the same time we need to recognize that the very technologies with which we are engaged are inherently, inherently autocratic command and control systems.
They're not democratic in the way that many spokespeople in Silicon Valley would argue, as they in fact have been for the last what, generation, about the development of social media platforms which were seen originally as being a great democratizing influence in our societies, only we found that actually they contribute towards fragmentation and an undermining of the liberal democratic consensus. So we are, are we rushing to embrace, do we have any choice but to embrace technologies that are inherently owned by a very small number of oligarchic corporations and be that in the very way the code is written, so to speak, they are authoritarian. At that point, I think I should stop and just hand over to Elina.
Elina Lange-Ionatamišvili
Thank you very much, it's an honor to be here. I wanted to highlight a couple of future trends which we identified in this report.
So, my first point is that the contest for the attention is moving from the visible front to the machine front, if we may call that. So, today information manipulation is aimed at people, but increasingly it will be aimed at the systems that decide what people see. So, if you like, we could call that the curation level, or curation layer, where retrieval, ranking, filtering, and summarization happens. And this is the layer where the human and machine meet or interact, so to speak. So the manipulation there happens in two ways.
The slow way is what we already are aware of: poisoning of the training data so that a model learns a distortion and carries it forward, but that is going to increase. And the faster way is seeding material that a model will pull in when it answers a question. But importantly, in both cases, the content will be produced for machine consumption and not for direct human persuasion. And this also has a consequence for detection. Our current instruments look for narratives, look for engagement, for coordinated behavior, authentic or inauthentic, and none of these signals are present here.
They're just not visible in this kind of situation. I would also add one further element. In the future, adversaries could build highly accurate digital replicas of population, or so-called synthetic audiences, and test their influence approach on them before deploying in the real world. So that requires several kinds of data: what people do, what people say, what they believe, and how they are connected. And some of this data is well protected, but much of it actually is not.
And this is something that can be misused not only by hostile foreign actors. So this is something we also have to keep in mind. My second point is about who owns this curation layer. As agentic systems become more mainstream, information will reach people in a highly personalized way, in ways that we probably can't even conceive at the moment. And this may lead to a greater atomization.
Sorry, yeah, atomization. Is that correct? Yes, atomization, sorry. So basically that means people living in parallel individualized realities with no common reference point. And you can imagine what that does to democratic deliberation.
So where there is no shared standard of evidence, the argument stops being about whether a claim is true or not. It becomes an argument about which system has the authority to decide what actually counts as evidence. And that infrastructure is owned by private companies, by individuals who operate at a suprastate level, if we may say so, and at a scale that no individual state can match, in Europe at least. The big players are also of course based outside of Europe, as Dr. Bold pointed out. So my third point concerns neurotechnologies that I'd like to raise.
Our findings show that neural data will become a commercial asset, and consumer devices that for now at a very rudimentary form can read brain activity, they are already being tested for the consumer market. And it shows the direction in which the tech market is developing, and the European Union has no dedicated framework to address that.
I will stop here because we have just 15 minutes, and probably I think it's prerogative of the chair to pass the floor. Thank you.
Nicolás Pascual de la Parte
Thank you, thank you very much to both of you for touching upon very, very interesting points. I think that here you have been pointing, Dr. Neville, to some dilemmas we have to confront. But I would say that we confront not only dilemmas but trilemmas at the same time. We have to focus our energy, our investment in developing the emerging technologies, whether it is artificial intelligence, quantum computing, robotics, the data on the cloud, biotechnology and so on and so forth. Second, at the same time, we have to regulate how best to accommodate this with human rights.
And third, how to keep pace with China and with the US, that you said that already the train has left the station. I don't agree so much with that, but anyway, we have to face this trilemma at the same time, keeping in mind one thing, that there are three different models when dealing with artificial intelligence and disruptive technologies. The Chinese model, which is to develop technologies to control their own people, to be intrusive in their private life and to control. Second, the American approach, to make profit as quick as possible and as much as possible, capitalistic approach. And the European approach, where we place at the center of our preoccupation the human being, the dignity, the freedom and the human rights of the, and these are our own features of the European approach.
We have not to compete with the Chinese in controlling our people or with the Americans in doing quick profit. We have to develop the new technologies keeping in mind that the human being and his dignity is at the center of our preoccupations. But anyway, I am not an expert, that's why I wanted to give the floor now to Dr. Halyna Padalko for 10 long minutes.
Halyna Padalko
Thank you, I would not mind if you pass the remaining one minute from Alina to me because I have lots of recommendations and I have lots of maybe not very pleasant data for all of us. So thank you first of all so much for elevating this topic because we all need to concentrate more and more attention on this area right now, especially these days, you see what is happening with agentic AI on the backdrop of our discussion today. So today I'm going to be speaking about three emergent major roles of AI at the intersection of information manipulation. First of all, AI already have been used widely as assistant. So apparently I'm talking about, can we change the slide?
So how should we operate with that? Yes, thank you so much. I'm going to be just saying that I ask for changing the slide. So first block and big function is AI as assistant, and all of you have been already seen plethora of AI-generated AI slop on the internet. Apparently it lowers the cost of production of text, of any type of content.
It drastically changed the business model of information manipulation. For instance, if previously Prigozhin and Internet Research Agency had to hire 100 people or even more to conduct foreign interference into US election 2016, now companies like him, like Social Design Agency, only need few people to actually coordinate massive information operations to convey messages in favor of Kremlin. Second is AI became as a battlefield itself, it's actually what Alina have been referring to. It's what we call LLM poisoning, and by LLM grooming I mean deliberately seeding information into a training data set to shape large language models' outputs.
And given that more and more services and applications that we use on a daily basis, especially AI assistants, online coach or online psychologist even, to whom citizens, journalists, policymakers are asking what is true or to whom they share all their weaknesses in a minute of loneliness, this function of AI and this emerging role of AI becomes more and more concerning. Because lots of people are lonely, they seek digital companionship and AI is able to provide that. And Russians are able to poison AI. And third is AI as enabler, and that's what we currently are talking more and more about. So by this I mean proliferation of agentic capabilities and agentic influence system we widely contemplate these days.
Can I ask audience how many of you are following the Hugging Face and OpenAI incident? Just raise your hand who knows what is happening. Amazing, thank you. So agents are capable of maintaining personas, monitor conversations, choose target by themselves, test messages, do AB testing, and all of that are partly or fully automated. So currently we are encountering one of the biggest scientific problem in our field: how to detect AI swarms, how to detect the behavioral of these AI-enabled systems and especially AI swarms.
OECD report on the future AI risk provides few lessons for us, if we can have another slide. Slide number three, please. So first of all, report identify information manipulation along with its intersection on democracy as priority risk, priority AI risk, and it treats it as institutional risks for democracy, that's what Dr. Bolt was widely talking about. Secondly, information manipulation is not in a vacuum. It's very deeply interconnected with other risks posed by emerging technologies, like malicious cyber activities, invasive surveillance, etc.
And third, response must be comprehensive and include different phases of responses of democratic countries. For UN Independent International Scientific Panel on AI, it's three warning for us too. So first of all, industrialized persuasion. Secondly, a warning that of synthetic consensus, that agent and artificial personas can infiltrate communities and manufacture the appearance that many independent people support the same position, so like this coherence synthetic. And the third is epistemic erosion and liar's dividend.
On the other hand, we talk a lot about big companies as our enemies, but they are also suffering extensively by threats from our adversarial countries and misuse and exploitation of the very Western models for their malicious goals. First of all, recent report, can we have another slide, please? Recent report, 25th of August, like this is literally last week. Everything is developing so fast in this area that if you leave for a week for a vacation, you are outside of the basically main main events happening in area. So OpenAI reported disrupting Russian-origin account promoted the International Burke Institute.
This is artificial fake think tank from Israel that contained copies of misattributed academic work and sovereignty index favorable to Russia. So the operation manufactured authority and infrastructure of credibility was a product of this operation.
Anthropic also have been noticing misuse of their models and documented influence as a service operation going beyond content generation. Its operation managed more than 100 bot accounts on X and Facebook that used Claude to maintain political personas, decide whether accounts should like, share, comment on or ignore real users' posts. So this is like the little AI swarm. So far Europe have invested heavily in exploiting falsehood, debunking narratives. Can we go for next slide?
Moderating content and discussing watermarking and labels, which is like very important and profound discussions that should and have happened. But we cannot address the core problem on its own. We need to think systematically and not chase the content. First of all, simply a lot of content. Leaked internal Social Design Agency, this is like Russian and Kremlin contractor, record claims that bots army generated 33.9 million comments in four months only in 2024.
Now the number is like extensively higher. Secondly, AI labels do not neutralize the influence because of our cognitive and confirmatory biases. And lastly, democratic authorities are not in a position to establish the truth. So what do I propose after all of that terrifying information I just throw on you? Can we have a last slide, almost last slide, second one, next one up, yes, policy recommendation.
So first of all, my first recommendation, I would really love to hear your feedback on that. Create a category of high-risk AI information influence system comparable in logic to Digital Services Act recognition of very large online platforms. Classification should depend on capabilities, scale, context and foreseeable impact. Relevant indicators include automated persuasion, synthetic identities, behavioral targeting, etc. Second, extend AI Act implementation and future review from generated outputs to agentic capabilities.
Providers should test for impersonation, multilingual propaganda, account operation, coordinated agents, AI swarms and other information manipulations. For instance, I would highly love to highlight in this area AI companionships, coaches and mental health systems that provide emotional vulnerability or provide persistence behavioral advice should not be deployed in a sensitive context unless providers demonstrate safeguards against covert external influence and political or commercial steering. Third, go after actors, not after content and disrupt their operational chain, map the connections among models, accounts, domains, advertising, payment system, data brokers and state sponsors. We need to have them all mapped. Fourth, enhance interoperability.
I know you might be sick of this word, but we do need to have interoperability among our trusted partners, allied partners in Europe. DISARM and other information manipulation frameworks are useful, but they don't always cover, for example, cybersecurity reporting and other type of threats. Fifth, use AI not only for defensive but also to disrupt information kill chain and go offensive with offensive information operation within adversarial countries. Europe should invest in multilingual analysis and AI for detection, but also know and understand adversarial structures, identify weak points of adversarial societies, disrupt hostile infrastructure and launch operations inside Russia. And the fifth one, sixth one, invest in research to build capacity and support educational efforts and digital literacy among the population.
But AI can automate decision, but it doesn't remove the human and political actors behind them. Even if we are talking about such thing as autonomous weapon, I'm sure you talk a lot about autonomous weapon in this auditorium and others. But let's take one example from a material world and kinetic war, for example. When, for instance, autonomous weapon strikes a civilian in Ukraine, responsibility cannot simply be transferred to the machine or algorithms that made that. It was the very Russians who designed, selected, deployed, or simply exploit Western models for their malicious purposes.
And the same applies to information warfare and systems that decide how to deceive European citizens. The system might select messages, target, build the program, but actor defines what we call in math objective function and supplies the infrastructure, while algorithm only performs specific tasks. Yes, the future information environment will be shaped by AI, undoubtedly. But responsibility for its malicious use will still remain with the actors who build, direct, and deploy it. So let's expose them, disrupt their kill chain, and go on offense, investing in deployment and development of our own system.
Thank you so much, I'm ready for your questions.
Vasile Dîncu
Thank you. Thank you very much to our experts for reports, very important reports for discussion and a very good management of time. I will now turn to members for the Q&A session. We'll start with coordinators of representative of political group of EUDS and SEDE committee. I also open the catch-the-eye.
For the beginning, Eszter Lakos from EPP EUDS. Please.
Thank you. Thank you, Chair, and thank you very much for this very insightful presentation. I'm not saying that I'm fully aware of the nitty-gritties, but it was extremely stimulating. We have seen, of course, how the information environment is manipulated with the help of AI.
Please slow down for translation.
Eszter Lakos
Ah, okay. I'm so sorry. Okay, so I will just go to the questions. First of all, it's the neural data that you have said that it's of course an asset and also that can read the brain in a rudimentary way already. And you also mentioned that there is no EU framework yet, but of course how I see it and with the very swift change of technology, it can be really a kind of neural warfare and we can easily find ourselves in a kind of 1984.
So I was wondering, you know, at stake of that, what you see that the next steps could be. And also to the third speaker, to Ms. Padalko, what do you see as key elements to go offensive on AI, if you can mention really the three things. And on poisoning data, if you could dwell a little bit on that, because as a consequence we are going to have misleading AI that we use, be it everyday business or not. What could be done on this? And I will save the last question.
Thank you. From EPP side, Mr. Nicolas Pascual de la Parte, please, Mr. Co-Chair.
Nicolás Pascual de la Parte
Thank you. Thank you, Mr. President. I think that when we talk about manipulation of disinformation, we should be careful, because the information manipulation has been ever been with us all the time since information was invented. It's only that before, the monopoly of this information was in the hands of the state, and now it's in the hands of many operators in the market. I prefer to have plurality of operators in the market than to be subject to a single monopoly of disinformation by the state.
That's what we had before. Then let's talk of, let's approach the problem with a very, a more optimistic viewpoint. I prefer to have different sources of disinformation, I will choose which one will disinform me, and not to receive the only one from the state. First of all. Second, we Europeans, we are open societies.
We cannot behave as the Chinese or as the Russians to combat, to combat disinformation campaigns. I mean, you must not be a cannibal to fight cannibals, you know? I mean, we are open society, we have to keep the DNA of our societies that we, I prefer to have a scope of being manipulated, of being cheated, than to be controlled by a Big Brother who tells me what to read, what to listen, and what to think. Then, in the fight against manipulation from other illiberal, autocratic powers like China, Russia, we have the risk of confronting them with their own weapons and to look like them, to look more Chinese, not the Chinese look more European, but we look more like the Chinese do. This is a risk we have to avoid at any price.
Otherwise we will lose our DNA of open society. Third idea, I think that in Europe we have already agreed upon a number of instruments. Here we have the FIMI, the FIMI toolbox, which was agreed upon. We have the Artificial Intelligence Act. We have, we have also the, the toolbox for cybersecurity, for hybrid threats.
We have a lot of toolboxes here in Europe. The problem is how to activate them quickly and effectively enough. And for that we need mainly two or three requirements. First, trust and confidence. We have to trust to share information, intelligence.
We don't. We share only open sources information, not delicate intelligence. Second, to have a quick decision-making process after collaborating, exchanging our information among ourselves. It cannot take months to the attribution of responsibility. It has to take days at the most.
And third, to have a robust package of sanctions against those who manipulate our, our information scope. Keeping in mind, and that's my last remark, that disinformation has different levels. I mean, there's a basic level, which is fact-checking. For instance, Napoleon, Napoleon invaded China at the beginning of the 19th century. Well, it is a very, it is a very basic disinformation because you have the possibility of fact-checking immediately.
It was not China, but Russia, and Napoleon invaded. But there is a higher degree of sophistication.
You can say, okay, Napoleon invaded Russia to free the slaves from the Tsar. Which is not, this is the wrong interpretation of the invasion of Napoleon of France. But even you can go higher and to admit that he went there for other purposes, but to interpret the interpretation. I mean, the sophistication of the interpretation of the information is something that you cannot control at 100%. And that's why I come back to my first remark.
Let's let the individual, individual at the end of the day to choose the source of information, to choose how to protect himself or herself, and not to concentrate in any political power such a decision. Thank you.
Thank you. The next speaker, Mr. Ruotolo.
Sandro Ruotolo
Grazie, Presidente. Ma devo dire che il titolo, le sfide del futuro panorama informativo, è molto complesso. Allora diciamo che il futuro dell'informazione dipenderà dalla nostra capacità di governare lo spazio digitale, dove oggi si gioca una parte decisiva della battaglia per la difesa delle democrazie liberali e per la libertà di stampa. Il punto non è soltanto che i giornali vendono sempre meno copie, mentre i social media hanno miliardi di utenti. Il punto è che è cambiata la struttura stessa del potere sull'informazione.
Quando parlavamo di libertà di stampa e pluralismo, ci ponevamo due domande fondamentali: chi possiede i mezzi di informazione e quanti di questi possono essere posseduti dallo stesso proprietario soggetto. Oggi non bastano più, perché tra chi produce un'informazione e chi la riceve è comparso un nuovo soggetto: la piattaforma. Il proprietario di un quotidiano controlla un giornale, il proprietario di una televisione controlla la televisione. Una grande piattaforma digitale controlla invece l'infrastruttura attraverso la quale circolano anche i contenuti prodotti dagli altri. Fake news, sistemi ibridi, tutto quello che volete.
L'algoritmo. Chi è che decide quante persone vedranno quel contenuto? L'algoritmo, che non è neutrale. Decide cosa viene raccomandato, cosa amplificato, cosa invece scompare. E da chi viene costruito quell'algoritmo?
Da un numero piccolissimo di soggetti privati, interessati solo ai propri interessi industriali, commerciali e anche politici. E sappiamo che spesso rabbia, paura, polarizzazione, provocazione e disinformazione viaggiano più velocemente della complessità. Il professor Bolt ci ha ricordato oggi un punto ancora più profondo. Queste tecnologie non sono necessariamente democratiche nella loro stessa architettura. Ha parlato di sistemi che persino nel modo in cui è scritto il codice possono essere autoritari.
Come giustamente ha affermato la signora Padalko, non possiamo limitarci a inseguire ogni singola fake news. Dobbiamo intervenire sugli attori che permettono alla disinformazione di essere prodotta e amplificata su scala industriale. Questa questione riguarda la democrazia. Gli attacchi ibridi, le fake news, come si trasmettono? Nel digitale, e dunque piattaforme, e dunque investire nella sicurezza digitale.
Abbiamo un problema a Houston. Dobbiamo rispondere a questa giungla con la sovranità digitale europea e la trasparenza degli algoritmi. L'Europa non può essere la colonia di Cina e America. Siamo tutti d'accordo. Grazie.
Csaba Dömötör
Grazie, Mr. Ruotolo. The next speaker is Dömötör from Patriots.
Elmondta, hogy az új technológiák olyan óriáscégekhez kötődhetnek, amelyek nem igazán jellemezhetőek demokratikus és átlátható működéssel. Ráadásul, tehetem hozzá, ezek a vállalatok oligopol helyzetben vannak, ha nem is monopol helyzetben. Hogy ez mennyire erős dilemma, jól jellemzi egy pár héttel ezelőtti találkozónk. Volt alkalmunk egyeztetni egy nagy meghatározó közösségi médiavállalat európai vezetőivel, akik elmondták, hogy 90%-ban az algoritmus hozza meg a tartalmakkal kapcsolatos döntéseket. Azt hiszem, hogy megnyugtatásunkra azt is elmondta, hogy ne aggódjunk, lejjebb csavarták a politikai jellegű tartalmakat az ő közösségi médiaplatformjukon.
Thank chair, let me speak in Hungarian, let me ask in Hungarian. Bolt úr egy nagyon fontos dilemmára tért ki az előadásában.
Tehát ez a példa is mutatja, hogy mennyire erős, mennyire fontos ez a dilemma, és igazából teljesen mindegy, hogy ember hozza meg a tartalomkorlátozó döntéseket vagy az algoritmus, ez végső soron korlátozza a demokráciát. Hogyha ebben a dilemmában, amely nem kicsi, az európai döntéshozók arra jutnának, hogy igenis ragaszkodnak a demokratikus működéshez és a nagy mértékű átláthatósághoz, akkor azt szeretném önöktől kérdezni, hogy önök szerint ennek milyen szabályozási és politikai eszközei vannak, úgy is kérdezhetném, hogy vannak-e konkrét javaslataik ezen a téren? Köszönöm.
From ECR, Seded domnu Poznak Rainis.
Reinis Pozņaks
Yeah, thank you chair. I have million questions, but I will not ask any of it, just comments because it's, I don't know, for many it sounds like we all gonna die. And yeah, because the one thing is technology and of course Europe's sovereignty and all that kind of things, it's important, but I think the crucial thing is not about having technology, but how Europe would use it even if we had those technologies. Because when we talk about disinformation and manipulations, we've been talking about that for I don't know how many years. Every year the same thing, we see it, we know it, and I kind of agree that we are free democracy and everybody can make disinformation, it's fine till the moment when you do the disinformation that cucumbers are more healthy than tomatoes just to sell more cucumbers.
But the situation those days is that we are under attack, it's a war. And this disinformation campaigns are managed with the one goal to destroy basically Europe. And if we will not, and I fully agree finally in the last presentation, that if Europe, even if will have the best technologies, will not proactively attack enemy, we will not win. Because it's impossible to win in just defense or try to debunk some disinformation, it's losing game. We have no chance.
And again, I want to say we have to understand that when we talk about disinformation and manipulation, we talk about war and about specific attack on the Europe, not just like disinformation or manipulation in general. So and that is the big difference. And again, fully agree, we need to find the ways also how to attack on those who attack us with manipulation and disinformation.
Thank you.
Thank you, Alexandra Geese from Greens.
Alexandra Geese
Yes, thank you very much. I have a couple of comments and then a few questions. I think first of all to the representatives of the NATO Strategic Communications Centre of Excellence, I think it's very important to analyze situations separately in order not to completely feel overwhelmed. You know, when I heard, listened to your first presentation, I felt a little bit overwhelmed, oh, there's nothing we can do. There actually are a lot of things we can do.
And we do have an issue that technologically in terms of social media, chips, data centers, AI, we have to do things in Europe in order to have sovereignty. But in the meantime, when we talk about the information ecosystem, there's a lot of things we can do. Because what we suffer right now is what Dr. Podalko called the liar's reward. That means we have algorithms that govern social media, which systematically reward hate, out-group hostility, hostility against other people, moral outrage, and which punish, that means reduce the reach of facts, of traditional media and of parties in the center of the political spectrum. And technically that is called engagement-based ranking.
This is the principle that is currently governing practically all social media. And that is a principle we can change very easily because we have the Digital Services Act in the European Union that says in case that public discourse, elections, fundamental rights are threatened, the European Commission can ask these platforms to take mitigating measures. And a mitigating measure here would be to change these algorithms to give people their choice back. And I very much agree with the chair. I don't want a state, I don't want a government, a minister of the interior to control our information.
I want the users, I want the people to be able to decide on social media what kind of content do they want. And I want them to choose that actively and not having content pushed into their timelines because currently some of these billionaires controlling these companies work very closely with governments. Mark Zuckerberg very explicitly said he's applying Donald Trump's agenda in the world. I mean, he says this very, very openly. And this is something we can very easily change.
So that's my first suggestion in terms of solution, maybe you want to comment on that. Second one, could we have social media only for real people? Because the agentic AI, the botnet threat is very real. I'm very skeptical about getting rid of anonymity. I think that is a problem for the victims.
But having proof of being a real person should be something that technological we should get there. So not your home address and the name and your phone number, absolutely not, not your ID. But being a real person, I think that could solve a number of problems. In the EU, the Commission is working on developing an app for age verification. That would also allow people to say, well, I'm a real person, I'm a citizen registered somewhere in the EU, I'm not telling you who I am, you cannot connect this to your advertising ID where you store all my private data that you shouldn't even have, but I'm a real person.
And would that help, that's a question also to Dr. Podalko, but also to the NATO experts, would that help is that a solution? And then to Dr. Podalko, a special thanks for outlining very concrete solutions and very concrete steps. I think that was extremely helpful. I have two questions. First of all, you said create a category for systemic risk AI-enabled influence systems.
Could you be a little bit more specific? And recently the Commission designed ChatGPT under the DSA, so under systemic risk and so on. I don't know if you're familiar with that piece of legislation. You mentioned the AI Act, but maybe you could outline a little bit more for me and colleagues what that would mean.
And a second question, OECD says liability. This House took the decision and the Commission took the decision not to have a common AI liability framework. We should have gone with the AI work. How do you see that today? You mentioned the Hugging Face case.
I think some colleagues are not familiar with it. What practically happened is that hundreds and thousands of AI agents teamed up to hack into a private company's systems, to make it very, very short. So if you do that as a private person and you hack yourself into a bank or a private company, that's a criminal offense. If you do that as a group of a thousand people, that's an even more severe criminal offense, it's probably a criminal association. If you do that within OpenAI and you are the CEO of OpenAI who's enabling and actively encouraging that, taking away the safeguards, you get invited to the G7.
I think that is a problem. So what can we do in terms of liability? Thank you.
Mārtiņš Staķis
Thank you so much. Yes, today we talk a lot about artificial intelligence and how it affects our societies, but my question is about another case. So last week, The Atlantic reported on a paid clipping operations run from Riga called Elephant Clipping. Basically what they do, they pay people outside the United States for every view they got on short pro-Republican videos. And these videos look like just normal people who just, you know, sharing their opinion, and nobody even know that this is paid.
And we have seen this playbook, right? It was actually Russian playbook and we saw this version of this at the Romanian presidential elections. And so my worry is that more and more our social media feeds get flooded with a paid content and nobody ever realize it is paid. And who knows, maybe one day it's one government, another day it's, you know, Russian government. So my question is simple and it goes to Ms. Elina Lange-Ionatamishvili.
Does the EU and national governments currently have the instruments to detect and fight these type of methods? And if we don't, what is still missing and what is the role of AI into this, yeah, to tackle this problem? Thank you.
Thank you. Pekatoveri from EPP, it's the floor, please, one minute.
Pekka Toveri
Thank you, Chair. Thank you for good briefings. I have to say that NATO Strategic Communications Center's briefing didn't help my sleep. On the other hand, I a little bit disagree about that we have already lost the game because if the investment bankers and economists would be always right, we wouldn't have this economic crisis every 15 to 20 years. So the game is still open.
I totally agree with Mr. Borrell Fontelles that yes, let's go offensive. You never, you don't win wars by defending. Let's go offensive, let's go after these, let's create capabilities. You know, the best thing of course would be the bomb on the bot farm, but let's not go there. There are other things we can do.
But one question, has, have we already lost the faith in the target, you know, the people? That the people would be, you know, the kids would learn to know and understand, have enough common sense and enough knowledge about history and society that they would immediately recognize that well, this can't be true. Or are we just talking to find technical means or lawfare to stop these attacks? You know, do you think, I'm a bit skeptical, I have to say, even in Finland where we tend to think that we are well educated, but we see that the younger generations are not getting any brighter, at least when it comes to social media, so that they would understand to be skeptical about these things. But do you think that there's something we still could do to harden the target, because target is the human?
Thank you.
Fabrice Leggeri
Merci Monsieur le Président, je vais m'exprimer en français. J'avais une question assez factuelle. Parmi les recommandations, il a été proposé, si j'ai bien compris, d'attaquer ou de détruire plutôt les réseaux hostiles ou les machines hostiles qui nous attaquent. Et donc je voulais savoir sous quel cadre juridique, est-ce qu'on est dans un cadre juridique plutôt militaire de défense, est-ce qu'on est dans un cadre civil, donc relevant plutôt, je dirais, de la coopération policière et judiciaire? Et puis concrètement, comment faire, puisque c'est une proposition qui a été entendue de la part de nos invités.
Merci beaucoup.
Merci. And the last speaker, Sousa Silva from EPP.
Hélder Sousa Silva
Thank you, thank you chair. First of all, I would like to welcome the experts and the researchers to the European Parliament and thank you for your presentation. Your scenarios close on a point I found particularly interesting: the compromising of computing and data processing capacity in orbit. I'm speaking about satellite communications. We are steadily moving processing off the ground and into space, and the information environment you described increasingly rides on that layer.
Much of our thinking so far has focused on the connectivity, and right now the European Union, we are building IRIS² for secure communications. However, and however sometimes it's the more important, the processing layer in space is newer ground. So two questions, if I may. On your timeline, is this a threat for the 2030s or one we should already be preparing for today? And second, much of this capacity is likely to be commercial and much of it non-European, and that's a key point for me.
And the question is the following: how comfortable should we be building our information resilience on orbital infrastructure that a private operator or a third country private operator, for example Mr. Musk, or a third country could ultimately restrict or switch off? Thank you so much.
Thank you all for the question. I now give the floor again to the speakers, starting with Dr. Boldt and Ms. Lange-Ionatamishvili. Please to limit the reply for maximum five minutes in total.
Neville Bolt
Thank you. That's fascinating listening to all the different opinions and different perspectives, because clearly there is no consensual understanding of A, the technologies we're dealing with, and B, the problems that we face. And that's good, that's healthy, that's to be celebrated.
But at the same time, it gives us enormous complications in trying to even answer these questions today. Because I suppose what we're talking about, always have to remember, and one or two people today have touched on that, which is timeframes. The chair originally said, you know, how quick can we make quick decision making? How soon can we, what is quick decision making in any hierarchical governmental kind of model? What constitutes fair, ethical, accountable and transparent decision making?
At the same time, they were trying to shorten, and I think quite rightly, shorten the periods of reaction. But even when I use that word reaction, I think there's a problem in that. Because we are, and a few people have mentioned that today, we are overly reactive in our mentality when what we need to be is preemptive. Some argue for disruptive attacks on those malign forces who threaten liberal democracies, but others at the same time quite rightly say, well, on what basis are we conducting the attack? So we heard one version, one account that said we must not descend to the level of the Chinese.
We must not fight a Chinese model with Chinese thinking. We're better than that. We're better than that. We believe in fundamental human dignity as the driving force of our societies. So I think we face, it's almost like we need far more people working, greater resources, greater assets and more people who are working on all the different aspects of what we're talking about today.
Those who are actually dealing with the immediate, the present. And yet we need people who are also working, if you like, in advance, who are preemptive. You know, here we're really talking so far just about what is, I think, a pretty basic technology with artificial intelligence. Alina touched on neural microtargeting, which I think is the real threat. But on the one hand, we need to be very sophisticated in our thinking, which is, if what technology companies are trying to do is not only capture the data from our emotions and let's call them just cognitive processes, rational processes, it's not just capturing the data, but clearly, and this is quite consistent with what we've seen in the last hundred years.
There is nothing unusual about the idea that governments, but also commercial organizations and at the sharp end of that of the commercial process, advertisers and marketers have, I mean, think about 1935, Gallup introduces the Gallup poll. It's not that he's trying to capture data, he's trying to predict human behavior. That's the point of it all. And this is, I think, where we need to be thinking with microtargeting using neural network technologies. That's, I think, where we should be focusing a lot of our attention, which is into preemptive thinking and also creating the right legal frameworks and also, where necessary, the right financial frameworks.
You jump in.
I'll try to do this in less than a minute, but I will perhaps first address a question by Mārtiņš Staķis, labdien.
Elina Lange-Ionatamišvili
I'm actually surprised that anyone is still being paid for that, that this is still a kind of a paid human activity by and large. I would expect very soon to see this completely automated. So the only investment would be setting up sufficient number of agents who conduct this kind of activity on the behalf of the operator. In terms of investigation, prosecution, yes, of course we have the capability and also largely relevant legislation in place to follow crypto, you know, any kind of organized money moving around by malign actors. However, that should not be us who spend time on this and chasing them.
It should be the platforms where this is happening and they're clearly not doing it sufficiently. So this is yet another thing to raise. I will probably just quickly to the first question about the neurotech. I'm not expert in European legislation, unfortunately, but for me the first obvious place where to look would be GDPR. So closing the gap between medical device rules and consumer device rules, what applies to which data that's being collected before the market decides for us.
So that would probably be the answer. But since we're out of time, I'll stop here. Thank you.
Thank you, Dr. Padalko. You have the floor for five minutes, please.
Halyna Padalko
Thank you so much. Thank you for your reflection and for your time. I appreciate that. It was very interesting to get to know what do you actually concerned with and I'm glad that discussion about social media is still going, but I think what we are contemplating widely, the researcher community, we do see more and more of exploitation of LLMs and AI as a game changer in this area. So lots of comments, I'll try to address everything and if I will not have time, we'll just chat in person later.
First of all, offensive actions will not make us Russians, especially if we admit that we are at war. And I'm glad a lot of people here do share this sentiment too. Being reactive and chasing every fake to debunking, as a lot of European agencies have already been doing brilliantly, it will not help us to win information warfare. And only proactive position with disruption of infrastructure, as I said. And this might not be always Mr. Leggeri like physical infrastructure.
However, we do have cases when Viginum, which is French actually authority, one of the best ones of that kind, was disrupting physically offices of a pro-Russian organization. But very often if we are talking about that kind of infrastructure, it is digital. So and it is our job to understand and map all of them, starting from Sergey Kiriyenko in Kremlin, finishing with Social Design Agency and other contractors and ending with agents who are in France, in Bulgaria and everywhere. So this is one of the biggest tasks that actually AI can also help us with to map all of that and understand the big picture. Secondly, on the offensive side, I hear you, a lot of you were interested in this particular part.
First of all, on the offensive side, we have been talking a lot on this democratic, pro-democratic messaging. And there was a brilliant question: have we lost the target, meaning that have we lost Russians? I do have my own personal opinion on that, but lots of efforts that our colleagues have been doing there, trying to convey democratic messages, trying to basically spread the truth there, were not really successful. Talking about offensive, I actually mean first of all being proactive with disruption of their infrastructure and understanding how infrastructure looks like, who are the actors, what are the main nodes on the picture. But also, what are the vulnerabilities and weak points in their system, the same way they do understand our societies and our weaknesses and try just to deepen polarization in our societies, because the main goal of Russians is not just to win information warfare, but undermine democracy as institution.
So we do need to understand with respect to the war that is happening, what are the weak points in Russian society. Is it Russian wives, pregnant Russian wives sitting somewhere on a block post and watching some certain type of information? We do need to understand that to convey our messaging there, that will help us to reach our kinetic goals on the war and also protect our democracy. If it's some specific type of Russian engineers that are producing ballistic missiles that are currently targeting people in Kyiv, we do need to understand where they hanging out, what are they reading and how to convey information to them and explain what they are doing. This should be very precise and weak points and weak targeted interventions.
And so far, we are not really doing very great with that regard. Okay, I'm going to make a last comment about something I think is very essential and critical. Yes, we are open democracies and that's something that actually make us different from Russia and other autocratic regimes. But unfortunately, I would not share the comment that we should allow our citizen to be disinformed in a way that they want to be disinformed, because that's actually what Russians would very much prefer to hear. On that note, I would say that we need to work more, I understand that this is security and defense committee, but if we were speaking for educational committee, I would say we need to develop and deploy into every school programs for digital literacy, information literacy and teach our kids how they can be misinformed or disinformed and what adversaries are actually doing.
I'm teaching a course at MIT and one of the classes we actually develop information campaigns with our students. And that's absolutely rewiring their brain, because after that they start thinking as adversaries. And every time they see malicious or manipulated content online, they are trying to question and reverse engineering, reverse engineer what they just saw. So I think education will save us and save democracy, so if we invest in offensive operation and education, that's something that will actually help us. Thank you.
Thank you all of us for the question and answer and for final please Mr. Nicolas Pascual de la Parte some conclusion.
Nicolás Pascual de la Parte
Thank you very much. I would dare not to draw conclusions of a very open debate and rich exchange of views. I think that it is only the first step in a long way of exchanges. I hope that it will be open, this space will be open in the future for incoming guests, experts from the civil society at large, from the political scope, the enterprise, the economy, the industry. And I think that little by little we can come to basic consensus.
It's clear that disinformation is an instrument, could be even an instrument of war, as Russia is using it. And that's why the most pressing need, as my colleagues were saying before, is that we have to prepare ourselves and to provide ourselves with the means to counteract this, countering this, this Russian disinformation campaign, knowing that it will be with us forever. Disinformation is a kind of challenge, multifaceted challenge which has come to stay with us forever. We are not going to wipe it out. Then, first of all, let's provide ourselves here in Europe, at the level of member states, but above all at the level of the European institutions, with all the means available to countering this disinformation campaign carried out by Putin's Russia.
Second, I said that also sharing the common feeling here is we have to be proactive, not only defensive, reactive. If you only defend yourself when you are attacked, you have lost half, 50%. Then we have to cobble up together our own narratives, our own interpretations, and to disseminate in Russia, in the Russian society, and in the other, in China and so on and so forth. We should have, we should be able to be an active global player also concerning the dissemination of information all over the world, including China, including Russia and the authoritarian regimes. Third idea, I think that that means that we have to be on top of the, of the technical and technological race, obviously.
We have to be in the cutting edge border of this. We cannot sleep away and we have to invest massively in these new disruptive technologies, whether it is artificial intelligence or quantum computing or 5G telecommunications, robotics and so on and so forth. And the last idea perhaps is that in addition to this technical and technological approach, we need a political discourse. A political discourse here in Europe, a protocol of reaction, a common protocol how to deal with these campaigns which are harassing us from Russia, from China and from other, other global players. We have to have a common sense of urgency, of purpose and of unity.
Thank you very much. And thank you to the special guests.
Thank you. We'll continue with EODS after two minutes.
Vasile Dîncu
Dear colleagues, please take a seat. We continue with exchange of view with Lubna Jaffery, Minister of Culture and Equality of Norway on the Norwegian strategy for strengthening resilience to disinformation. This our next point is Ms. Lubna Jaffery, Minister of Culture and Equality of Norway, who is joining us online. Minister, thank you for accepting our invitation today. Minister Jaffery will present the perspective of a country located in the European continent with shared common history and value with EU member state.
We also share common adversary. The Norwegian strategy for strengthening resilience to disinformation identify Russia and China as primary drivers of disinformation campaign. This is a well-through-planned strategy stretching between 2025 and 2030 with a prominent role for education and culture as levers to encourage citizen to distinguish between trustworthy and false claims. We have many question but I try to pronounce two. EU legislation do not automatically apply to Norway.
However, you are looking closely at the Digital Service Act. What is the current state of play regarding its incorporation into European Economic Area agreement and its transposition into Norwegian law? What do you see as to the main obstacle for implementation this EU legislation? Two, the strategy mention a role for libraries. How do you envisage engaging library in a world where citizen heavily rely on online sources?
And finally, schools are part of your plan.
Has disinformation been incorporated into school programs? Do you foresee for use and gamification to better educate children? Minister, thank you for being with us. The floor is yours for 10 minutes.
Lubna Jaffery
Thank you so much and Chair, honorable members, thank you for the invitation. It is a pleasure to be here today. Though Norway is not a member country, we have been a close partner since the European Economic Agreement was concluded in 1995. We share borders through the Schengen Agreement, we share the same democratic values, and we share many of the same challenges, including to our shared information environment. I therefore welcome this opportunity to exchange experiences and perspectives with you on how we can strengthen our resilience against disinformation.
As you all know, this challenge is not new. Disinformation is not new. What is new is the scale, speed, and sophistication. Today, social media and digital platforms allow information to spread across borders in seconds. Generative artificial intelligence has made it cheaper and easier than ever to create convincing text, images, audio, and video.
At the same time, the security situation in Europe has changed significantly. Authoritarian states use influence operations and disinformation as tools to weaken democratic societies, create division, and undermine trust. This is not only a security challenge, it is also a democratic challenge. Democracy depends on citizens being able to access reliable information, evaluate competing claims, and participate in informed public debate. If citizens no longer know what to trust, democratic decision-making becomes more difficult.
No country can meet this challenge alone. The platforms are global, influence operations cross borders, the technologies are largely developed outside our national jurisdictions. That is why cooperation between European countries is essential. Norway strongly supports European efforts to promote transparency, accountability, and democratic oversight of the largest online platforms. The Digital Services Act, the European Media Freedom Act, and other relevant initiatives provide important tools for addressing systemic risks while safeguarding freedom of expression.
We also welcome the European Democracy Shield and will follow the work of the new European Center for Democratic Resilience with great interest. In this area, there is a great value in learning from one another, exchanging information, and coordinating our efforts. Next slide, please. When it comes to information manipulation, Norway starts from a position of strength. We have high level of trust, we have a highly educated population, we have strong protections for freedom of expression, and we have a diverse and independent media landscape.
Trust in editor-controlled media remains high compared to many other countries. But we are not immune. Like other democracies, we see changing media habits. Artificial intelligence is rapidly changing how information is created, distributed, and consumed.
Our approach to the problem of information manipulation is rooted in Article 100 of the Norwegian Constitution. The article guarantees freedom of expression, but it also includes what we in Norway call the infrastructure requirement. It states that the authorities shall create conditions that facilitate an open and enlightened public discourse. This principle has been central in our work. It means that the state's role is not to decide what people should think, nor is it to eliminate all false statements from public debate.
Our role is to help maintain the conditions that allow citizens to form their own opinions and participate in democratic life. In 2025, the Norwegian government presented Norway's first strategy for strengthening resilience to disinformation. The strategy covers the period 2025 to 2030. It is important to understand what the strategy is and what it is not. The strategy is primarily about prevention.
It focuses on strengthening resilience and thus reducing any potential harm from information manipulation. It is not a strategy for identifying and responding to individual disinformation campaigns. Those issues belong largely to other parts of government, including the security and intelligence authorities. The strategy is built around five priority areas. Next slide, please.
First, strengthening critical media literacy. Citizens need the knowledge and skill to access information critically. This includes understanding sources, it includes understanding algorithms, and increasingly, it includes understanding how artificial intelligence works. Schools already play an important role, but schools cannot do this work alone. Libraries, civil society organizations, researchers, and media organizations are also essential partners.
As part of the strategy, the government has supported a new independent center of source awareness. The center is called Think. This will develop educational resources and workshops in collaboration with researchers, educators, journalists, and civil society organizations. The aim is to strengthen critical thinking and source awareness both in schools and among other groups, including older citizens. Next slide, please.
Holding social media platforms accountable is the second thing. Social media platforms have immense influence over public discourse. Their recommendation systems affect what information people see. Their design choices affect how information spreads. The strategy therefore emphasizes better oversight, greater transparency, and stronger accountability.
Norway is preparing for the implementation and enforcement of relevant European digital regulations, including the Digital Services Act. Next slide, please. Third, strengthening editor-controlled media. Independent journalism remains one of the best defenses against information manipulation. Editorial media operate according to professional standards, ethical guidelines, and editorial responsibility.
They verify information, they provide context, they give room for different voices and perspectives, and they contribute to informed public debate. The strategy therefore emphasizes media diversity, local journalism, investigative journalism, and sustainable conditions for independent media through media support schemes. Next slide, please. Fourth, strengthening knowledge and research. We need more knowledge.
We need better evidence about how disinformation spreads. We need to understand who is most vulnerable.
And we need to know which measures actually work. And as a follow-up, we have therefore recently established a new research center for long-term research on public discourse. This work would include research on disinformation, polarization, freedom of expression, and democratic participation. Next slide, please. Fifth, strengthening coordination across government.
Resilience requires cooperation. Many ministries and agencies have responsibilities related to disinformation. The strategy therefore seeks to improve coordination, information sharing, and situational awareness across government. An important part of this work is expanding the role of the Norwegian Media Authority, which will have a stronger responsibility related to resilience, platforms oversight, and knowledge development. International cooperation is a cross-cutting theme through the strategy.
Through the EEA and Norway Grants, resilience to disinformation and information integrity have become important priorities in our cooperation with our partner countries. During this eight-year term, over three billion euros will be granted to 15 EU member states in Central and Southern Europe. This includes substantial support for critical media literacy, independent media, democracy participation, and responsible digital technologies. In Poland, for example, funding has been allocated specifically to efforts aimed on countering disinformation and strengthening information integrity. Moreover, the program seeks to strengthen civil society, promote gender equality, and LGBTQ+ rights and provide equal accessibility and opportunities for all.
These programs allow us not only to contribute but also to learn from our European partners and share experiences. Norway also has very good experience from cross-European cooperation exchange through the culture and media program Creative Europe. We are currently assessing the possibility to participate in the forthcoming Agora EU program. Hence, we are following the current negotiations on the EU multiannual financial framework with keen interest. Next slide, please.
Allow me to conclude with three brief reflections. First, resilience is not something we can legislate into existence. It depends on trust, it depends on education, and it depends on strong democratic institutions. Second, we must protect democracy without undermining the values we seek to defend. Measures against disinformation and information manipulation must always be balanced against freedom of expression, freedom of information, and democratic participation.
A few months after launching our disinformation strategy, the government therefore presented a national strategy for open and enlightened public discourse. While the disinformation strategy addresses a specific threat to democratic debate, the second strategy provides the broader principle-based framework for the government's work on freedom of expression. It reflects a simple conviction: our task is not to control public debate but to strengthen the conditions that allow it to remain open, informed, and inclusive. Third, European cooperation will become even more important in the years ahead. We need stronger information sharing, we need more research cooperation, we need better insight into platform risks.
And we need continued dialogue on how democracies can protect open and enlightened public discourses in an increasingly complex information environment. Our goal is not to eliminate disagreement; democracy requires disagreement. Our goal is not censorship; our goal is to ensure that citizens can participate in public debate based on facts, knowledge, and free exchange of ideas. Thank you, and I look forward to your questions.
Thank you, Minister, for excellent presentation. We pass to the question time starting with coordinators and political group representative. I also open a list for catch-the-eye. For beginning, Eszter Lakos from EPP.
Eszter Lakos
Thank you, Chair. Thank you, Minister, for this very impressive presentation on your strategy on disinformation. I found several points that they are very challenging and where we can learn from for our future also at national level. I would like to mention to you that the model that you have introduced is based on trust and I was, you know, wondering that there could be situation when the state media, for instance, can be captured. There are situations when the state has immense resources.
So we are facing the DIMI, so the domestic information manipulation and interference. And I was wondering what could be your take at this, how whether you think that DIMI is at the same threat level as FIMI, so foreign information manipulation and interference is, or you see that a different approach should be taken on these two aspects of info manipulation. And additional lighter question, but I was very amazed that libraries are your essential partners. If you very briefly could just describe what they are doing because I think media literacy, digital literacy has to be close to citizens. And as you said that your strategy covers the the youngs and also the seniors as well.
So what are your experiences that you could highlight in this regard? And thank you again.
Sandro Ruotolo
Grazie, Presidente, e grazie, signora Ministra. Lei ha detto una cosa che considero fondamentale: la disinformazione non è nuova, ciò che è cambiato sono la scala, la velocità e la sofisticatezza con cui oggi può essere prodotta e diffusa. E ha indicato due risposte che credo debbano necessariamente andare insieme: da una parte rendere i cittadini più consapevoli attraverso la scuola, la cultura, le biblioteche e l'alfabetizzazione mediatica; dall'altra rendere le piattaforme responsabili. Ecco, vorrei concentrarmi su questo secondo punto, perché possiamo insegnare ai cittadini a riconoscere la disinformazione, ma se gli algoritmi continuano a premiare i contenuti che generano più reazioni, polarizzazione e conflitto, rischiamo di chiedere ai cittadini di difendersi da soli da un sistema progettato per influenzare ciò che vedono. Lei ha parlato di maggiore trasparenza, controllo e responsabilità delle piattaforme e del rafforzamento del ruolo dell'autorità norvegese dei media.
La mia domanda è molto concreta: quali strumenti ritenete più efficaci per verificare se le piattaforme stanno realmente riducendo questi rischi?
E soprattutto le autorità pubbliche dispongono oggi di un accesso sufficiente ai dati e ai sistemi di raccomandazione per capire come gli algoritmi influenzano il dibattito pubblico? Conosciamo il codice dell'algoritmo? Perché credo che questa sia la questione centrale anche per l'Europa, l'Unione Europea. Non possiamo difendere lo spazio democratico senza sapere come viene organizzato e amplificato lo spazio informativo nel quale i cittadini formano le proprie opinioni. Grazie.
Thank you. Monsieur Fabrice Leggeri.
Fabrice Leggeri
Thank you, Chair. I will speak in French. Merci Madame la Ministre pour votre présentation. Vous avez fort justement indiqué que dans la lutte contre les manipulations, il faut trouver un équilibre pour garantir toujours la liberté d'expression et éviter des phénomènes de censure. Donc je souhaiterais sur ce point que vous puissiez nous partager votre expérience en Norvège, peut-être nous citer quelques exemples de difficultés concrètes auxquelles vous vous heurtez puisque les plateformes fonctionnent évidemment très rapidement, il y a une nécessité d'agir rapidement et en même temps il faut garantir la liberté d'expression.
Donc sur ce point, si vous pouviez nous donner davantage d'informations pratiques, s'il vous plaît. Et puis j'ai noté que vous avez évoqué la création d'un centre donc que vous souhaitez être indépendant, le centre Think. Et je voulais savoir là aussi comment le gouvernement norvégien, les autorités publiques norvégiennes s'assurent de la véritable indépendance, notamment par rapport au financement de ce centre. Et puis dernière question, mais qui est un petit peu dans le même ordre par rapport à l'indépendance réelle, vous avez évoqué un programme qui a pour objectif de financer la société civile. Et là aussi je voulais savoir comment le gouvernement norvégien, en tout cas le ministère qui gère ce programme, procède pour s'assurer que les associations ou les organisations de la société civile peuvent être indépendantes et vont faire un bon usage des budgets publics qui lui sont accordés.
Merci beaucoup.
Alexandra Geese
Thank you very much, Chair. Minister, first of all, I would really thank you for the great role that Norway has been playing in implementing some of the European digital legislation. Norway has been spearheading the fight for freedom, for example, also with the case brought about by the consumer associations against Meta on behavior-based advertising, which is another form of surveillance and control of the population. You said citizens need to understand algorithms and we need an open, informed, and inclusive debate, and this is something I very much share and I think most colleagues here share. But this is not what we have today because today this debate is controlled by the platforms, by the platforms' owners, which implement algorithms that very systematically reward hate, outgroup hostility, moral outrage, and which systematically downranks and punishes facts, traditional media, and parties which are not extremist.
This system is called engagement-based ranking, and we had a speaker in the previous session calling that the liar's dividend.
So do you think that the EU, which has the legal instruments with Article 34, 35 of the DSA, should use it in order to make sure that we change this specific algorithm and that we give people and citizens the choice back over their timelines, over the content they want to see? And do you think that Norway might want to play an important role in this fight? Thank you.
Thank you. From KDH, Elena Dupont from EPP.
Lena Düpont
Thanks, Chair. And thanks to you, Madam Minister, for your contribution and your insightful remarks that you also mentioned in the European cooperation, but of course also on the national level and the instruments you have had at hand. I think you mentioned one point that I find specifically, well, interesting, but at the same time also very complicated because it very goes very much down to the very essence of a democracy where you say that you cannot force people to like the democracy, cannot force people to participate in a democracy, but you very much rely on people trusting institutions and being part of a lively public debate. So I was wondering, in your first part of the presentation, you also mentioned that you are promoting of course that debate, but you're also addressing schools with that. So can you probably give us a bit of an idea beside the library, which I also very much like, likewise the comments from my colleague next to my side, because I think they are the excellent areas for freedom of thinking and education as such.
So what kind of instruments do you use and how to make sure that the results are also then indeed a very public and lively debate? You mentioned that you also constituted a new body, which I think was called Think or something. Maybe you can say a word on that in more detail. And last point on the editorial standards, and I finish with that. Can you give us a bit of an insight on how it is defined within Norwegian law?
For example, let's say the profession as such, is there a specific training for it? Are there certain standards that need to be fulfilled, for example, to become a journalist and so on and so forth? Thanks a lot.
Vasile Dîncu
Thanks for the question. Minister, please, you have the time for answer, complete answer.
Lubna Jaffery
Thank you so much. And thank you for a lot of questions, a lot of engaging questions. I will try my best to answer all of them. So on the issue of how we look upon threats in Norway when it comes to information manipulation, this is not a threat we have seen in Norway. We have had elections.
It's been a discussion if there will be information manipulation during the elections. We had election last year. There are no proof of that. And we see that the high level of trust we see here in Norway, and I know it will be maybe it's difficult for you probably to understand when I talk about that because it's something I think people need to experience. So you're all welcome to Norway.
But I think in large, this is one of the really, really core values of the Norwegian society, that we do trust each other. We do not, you know, think that the government is lying to me. You know, the pandemic is a very good example of that, how we had a lot of information coming, but there was also of course disinformation.
So someone, one of you asked me a question about do you have any examples of the freedom of speech and how you balance that in the discussion of disinformation? So I would like to use the example of the pandemic. So a lot of people, well almost, yeah, I can't remember how many people took the vaccine here in Norway, but it's well above 90%. It's a very high degree of vaccination here in Norway. But there were some groups spreading information on Facebook and other places that the vaccine is very dangerous.
We have had examples of people being severe ill after they're taking the vaccine and we also have people that have died. These are facts and this is true. But people have used those very, very small incidents and spreading, you know, disinformation about that the vaccine is very dangerous. So this is an example that people are allowed to say this, they're allowed to mean this and we as a society will not censor them. But we will go into a debate with them and have a discussion with them about this.
This is an example of how we're not going to censor people and we're not going to take down information that people are spreading on their own private profiles for instance. But when it comes to platforms, we need to hold them accountable. And the way we can do this, we can't do that alone here in Norway. Therefore I'm here together with you today because this is something we have to do together on a European level. So the Digital Service Act was also mentioned by the chair in the beginning of my introduction.
So the DSA, I think in my opinion, my colleagues, we have several colleagues in the government working with these issues. The Minister of Digitalization is in charge of the implementation of DSA. We are very eager in Norway to do that. I think that is one of the most important things we can do right now. So it is a negotiation between Norway and EU, so I can't go into the negotiation, but hopefully we will sort all the discussions we are having right now out and we will also be very, very eager to present the implementation of the Digital Service Act for the Norwegian Parliament.
So and when it comes to the algorithms, how they disfavor editorial-led media for instance when it comes to information, we see this very much in Norway. This is a huge discussion. I have had several meetings with the media platforms, with Meta, with Snapchat, with TikTok, they have been in my office. I also had editorial-led editors coming to my office having a discussion with them about these things. That we have been very, very frank with the platforms and told them that this is very, very difficult because this is a discussion about the freedom of expression and also giving people the possibility to get right information.
That influencers have more attention on the platforms than editorial-led media. So of course the editorial-led media can say we're not going to use the platforms because we are not getting enough attention. But they still want to be there because they see it as a democratic problem if they're not there to give the other people who are making their own news or just giving their own speeches not an alternative to that. So this is something we have had a huge debate with the platforms here in Norway on a national level, but to the question if Norway want to be in the seat and taking a lead on this, very much we would like to do that. There have been several discussions, questions about Think, yes, it's called Think, this center about, sorry, source awareness, the center of source awareness we have established here in Norway.
So yes, they are government funded, but also all the media in Norway is government funded. But we have a principle, the arms-length principle is very important. Therefore also we have a very high degree of trust that the government doesn't interfere with how they're doing their work. Of course they have to report on how they're doing their work, but we are not going to tell them you have to do research on this subject or research on this subject. They are free to do the work how they want to do the work.
So they are funded by the Norwegian state, but it is established by the media, it's established by publishers and it's also established by NTNU, which is a university in Trondheim here in Norway. So they are doing this together and they're doing it independently, also like the media is doing the news coverage, even though some of the media do get support from the media schemes here in Norway, they are very critical against media, no, against the Norwegian politicians. They're not afraid to ask us critical questions. And also the civil society in Norway, we have a long history of supporting civil society in different ways. But also the civil society, the arm-length principle is very, very important.
We do not mess with what they're doing and how they're doing their work. They are allowed to ask us critical questions. They're also allowed to put pressure on us to change certain parts of the Norwegian society. But we are very independent, we are dependent on them doing their part of the work because also in Norway you have the state, you have the educational institutions, but you also have the civil society that is a powerhouse in increasing and keeping the trust at a high level in Norway. So therefore we are very eager to trust also the civil society.
When it comes to media, no, sorry, when it comes to libraries and literacy. So libraries are a very core part of the Norwegian society. They are free. We have our own law, a public library law in Norway, that means that all municipalities in Norway by law needs to have a library in their municipality, which is free and open to the public. So this is therefore a very, very easy way of using the libraries in this task on challenging disinformation and making people aware of how they know this is disinformation.
And we also have school libraries that are very important. So I will focus on, we have two different very, very important action plans, or strategy, sorry. We have a strategy on reading amongst young people, which was launched in 2024. I think you can find it on the internet because we see the same challenge here in Norway as we see in a lot of other European countries, that the decline in reading is significant in young children. But we also see the decline now in adults and in young adults.
Therefore we're making a new strategy on reading also in the grown-up population because we need them to read so they can also be a part of the society, be a part of the debate and also get information that they can use to get informed and to use to get engaged and also use their voice during elections and also being a part of the civil society. Therefore the reading is very important and also the libraries and also the school libraries are very, very important part of that. Did I answer all the questions? I'm not sure. I think I covered most of that.
And if not, please tell me. Thank you so much.
Vasile Dîncu
Thank you, Minister, for spend your time with us. Your strategy, it's a very good example for us, for EU. We appreciate your vision, long-term vision, and your link democracy and education.
It's very important. I try to resume your lesson for us was resilience cannot be built with only regulatory and technology. It's necessary to build the trust and institution and the education is very important. Thank you, Minister, for your contribution and your important presentation. The next point is in camera.
I kindly ask everyone who is not allowed to leave this room promptly.