Reporting from the informal meeting between EU heads of state and government on November 13
The whole debate at riksdagen.se, in Swedish (opens in a new tab)
Translated from Swedish by AI; the translation may contain errors. The Swedish text is the original.
Speakers
EU AI development, regulation and competitiveness debate AI
The debate concerned AI development, regulatory frameworks, and Europe's competitiveness. M argues that Europe must participate in security negotiations 1 and that AI is a tool for competitiveness 1. M emphasizes that Sweden is a role model within tech 2 and that they have worked to ensure that the EU does not regulate AI in a way that harms innovation 2. M wants to invest in a Swedish language model 3 and argues that Sweden and Finland should have control over their own natural resources 4. S believes that AI development needs broad popular anchoring 5 and requests a coherent data center policy 5. SD argues that the EU's regulatory framework risks becoming a brake for small companies 6. MP argues that AI can reproduce society's biases 7 8. S believes that workers should not pay the price for the AI transition alone 9 10.
Written by AI in advance and may contain errors. The numbers lead to the speech a statement builds on; check against the text below.
Speakers (45)
- Statsministern Ulf Kristersson (M)
- Daniel Vencu Öhrlund Castro (S)
- Statsministern Ulf Kristersson (M)
- Daniel Vencu Öhrlund Castro (S)
- Statsministern Ulf Kristersson (M)
- Tomas Brandberg (SD)
- Statsministern Ulf Kristersson (M)
- Tomas Brandberg (SD)
- Statsministern Ulf Kristersson (M)
- Banesa Martinez Roldán (V)
- Statsministern Ulf Kristersson (M)
- Banesa Martinez Roldán (V)
- Statsministern Ulf Kristersson (M)
- Anna Lasses (C)
- Statsministern Ulf Kristersson (M)
- Anna Lasses (C)
- Statsministern Ulf Kristersson (M)
- Rebecka Le Moine (MP)
- Statsministern Ulf Kristersson (M)
- Rebecka Le Moine (MP)
- Statsministern Ulf Kristersson (M)
- Erik Ottoson (M)
- Statsministern Ulf Kristersson (M)
- Erik Ottoson (M)
- Statsministern Ulf Kristersson (M)
- Magnus Berntsson (KD)
- Statsministern Ulf Kristersson (M)
- Magnus Berntsson (KD)
- Statsministern Ulf Kristersson (M)
- Louise Minnhagen (L)
- Statsministern Ulf Kristersson (M)
- Louise Minnhagen (L)
- Statsministern Ulf Kristersson (M)
- Gurgin Bakircioglu (MP)
- Statsministern Ulf Kristersson (M)
- Gurgin Bakircioglu (MP)
- Statsministern Ulf Kristersson (M)
- Håkan Svenneling (V)
- Statsministern Ulf Kristersson (M)
- Håkan Svenneling (V)
- Statsministern Ulf Kristersson (M)
- Robert Olesen (S)
- Statsministern Ulf Kristersson (M)
- Robert Olesen (S)
- Statsministern Ulf Kristersson (M)
Statsministern Ulf Kristersson (M)
Madam Speaker! Last Friday, the EU's heads of state and government met for an informal summit in Dublin. Ireland is, as you know, the EU's presiding country at the moment.
The focus this time was on AI issues and certain Ukraine issues. The day before the summit, the countries also met in the broader circle of the European Political Community, EPC, with neighborhood and so on. In the margins of the meeting, the EU's long-term budget, support for Ukraine, and Europe's competitiveness were also discussed.
Even though the original intention was for AI to be the only, and important, point of discussion at the informal summit, several colleagues had requested that some specific Ukraine issues should also be discussed. This concerned partly who should represent Europe in any upcoming talks between Russia and Ukraine, and partly how the EU should secure Ukraine's financial support next year.
When it comes to various talks with Russia, President Putin has not yet shown any interest whatsoever in actual negotiations. If talks are to be held at all despite this, they must be based on the obvious: nothing about Ukraine without Ukraine, nothing about Europe without Europe, and Europe must speak with a common voice.
That is why it is fundamentally good that E3, as they say, that is Germany, France and the United Kingdom – two countries in the EU and one country outside the EU – were present and informally represented Europe in the talks that took place in Abu Dhabi last week.
In the discussions at the summit, we were in complete agreement on the fundamental point: Europe must be part of all conversations and negotiations concerning security – what is often called the security architecture – on our continent, and if it concerns the expansion of Nato or other Nato issues, it is only Nato and Nato countries that can make these decisions – not the EU.
Broad appreciation was shown for the E3 countries' work so far, and after discussions there was also a consensus that this format should be formalized and expanded to an E5 that also includes Italy and Poland. Representatives for the EU and Nato will also be included.
From a Swedish standpoint, this is not a perfect solution, but it is a sufficiently good model. No single person, after all, can be given a mandate to speak for all European countries within and outside the EU, and we do not want a cacophony of individual countries seeing their own, bilateral, reasons to speak with President Putin.
With that said, I want to repeat the importance of focusing on what is most important: to support Ukraine before and through a difficult winter and to increase our pressure on Russia. Dialogue can be good, but action is better. Most urgently now, the EU needs to secure financial support for Ukraine for next year. That question will be raised at the extraordinary meeting taking place as early as next week.
AI was also discussed, as said, quite a lot. To provide inspiration for the discussions, two guests had been invited to initiate them. It was the CEO of the French AI company Mistral, Arthur Mensch, whom one often meets right here in Stockholm. It was also the researcher and Nobel laureate Geoffrey Hinton, who is often called the godfather of AI.
The discussions had two main directions. The first was AI as a powerful European tool for competitiveness. There is a broad consensus on the importance of Europe not giving a technological walkover to the USA and China when it comes to AI.
We can enter such discussions from the Swedish side with great self-confidence. There are several prominent Swedish AI companies and an entire ecosystem of talent and capital that allow us to punch above our own weight. This was also noted at the summit.
The second part of the discussion was, on the contrary, about risks and how AI affects our societies in a much broader sense. For example, it concerned concerns about how AI affects children and young people, about national security, and about the long-term labor market.
It was also about what consequences the AI expansion, not least through the establishment of large data centers, has for our environment, our nature, and our energy prices. We are probably all in agreement on the importance of this expansion taking place in a responsible manner, and Svenska kraftnät is currently investigating, on behalf of Sweden, how large electricity connections can be managed in a more efficient way.
As is so often requested in AI discussions, global agreements were sought. Inspiration is drawn from other sectors, such as civil aviation, nuclear power, the central banks' financial cooperation against financial crises, and nuclear disarmament. There, even countries that stand against each other in many other issues can see the value of respected international regulatory frameworks. However, the way there is still long.
Daniel Vencu Öhrlund Castro (S)
Madam Speaker! Thank you, Prime Minister, for the report! I want to start by saying that it is good that the EU resolutely stands behind Ukraine and that we continue to strengthen our support for the country's struggle for freedom.
I also think it is good that the EU takes common steps to strengthen AI development. As the Prime Minister was touching upon, we in Sweden can lead in a direction because we have experience in applying new innovations and are generally positive towards new technology. To be positive towards new technology, Madam Speaker, is not a paramount goal. We are it because we see the benefits it provides for our citizens.
AI differs slightly from other technological leaps because AI has such an enormous impact on our democracies. Many experience that power is shifting to a few companies. The power over the labor market, our energy systems, and the information that comes to our citizens must be democratically anchored.
That is why we Social Democrats highlighted the concern regarding, for example, the establishment of data centers, something we noted that the Prime Minister also did before the council meeting. I would like to know if there is anything more concrete regarding a coherent data center policy in the EU.
AI development needs to take place with broad popular anchoring in order to strengthen trust in the development and respond to the anxiety that many feel regarding their place on the labor market, their electricity bills, and the increasingly polarized social climate.
Madam Speaker! I therefore think it is somewhat remarkable that the Prime Minister did not stand as a signatory to the statement from Finland's President Alexander Stubb and Norway's Prime Minister Jonas Gahr Støre on managing the risks of AI. I think, in fact, that Sweden should be involved in and leading the work for a responsible and competitive AI policy that delivers jobs and growth in strong democracies.
I would therefore like to know what was concretely concluded in the issue of data centers and whether the Prime Minister highlighted the need for more democratic control of AI development.
Statsministern Ulf Kristersson (M)
Madam Speaker! These are important issues. I will begin with the practical but important question of priority in electricity use. There are no European regulations here other than that one naturally does not have the right to favor national companies at the expense of Europe.
I mentioned that Svenska kraftnät, on the Swedish side, is working on finding a model where some kind of social benefit is seen in the order of priority. It cannot be a free-for-all to use very large amounts of power in the Swedish power grid based on pure queuing time. That work is ongoing. Several countries have mentioned the same thing.
This is to some extent set against the need to do things in Europe, which many also mention. A lot is being built in the USA now. There are probably strong reasons for European companies not to be completely dependent on computing power in the USA. These are classic trade-offs to make. Do we want to have the capacity in Europe, in the form of companies and clusters of powerful computers, for security reasons, commercial reasons, and privacy reasons, or do we want to outsource it to other countries?
Many European countries will need to make trade-offs. If we can do it together in Europe, it is better. That is essentially what Mistral's CEO and other CEOs of AI companies are calling for. We need power if the European companies, including the Swedish ones, are to be able to compete with the American and Chinese – but we are not prepared to do anything for this.
The democratic element is very much about creating international regulatory frameworks. I believe everyone realizes the need for regulatory frameworks and that everyone also realizes how difficult it is to get there. Uniting the USA and China regarding regulatory frameworks for AI is currently easier said than done, but it still does not prevent that discussion – on good grounds – from arising.
Daniel Vencu Öhrlund Castro (S)
Madam Speaker! We are, of course, also following Svenska kraftnät's work regarding data center establishments.
I note that Norway and Finland are now going further. They are introducing rules for the grid connection of data centers and also have a developed data center strategy, which I think we in Sweden can draw inspiration from.
My question linked to data centers is that many European countries are facing such establishments. It is of course so that when we have a common energy market, we can have a more coherent data center policy at the EU level. I would very much like to hear about the Prime Minister's view of this, because in the USA – which was also mentioned – data centers account for 4 percent of electricity consumption. It is an enormous amount of electricity that is being utilized.
Madam Speaker! I want a bit more concretization regarding democratic control. Why did the undersigned Prime Minister not sign the statement from Alexander Stubb and Jonas Gahr Støre on the issue that I believe the government – regardless of whether it is a transitional government or not – has broad support for in the Swedish Riksdag and which concerned more democratic control over AI development?
Statsministern Ulf Kristersson (M)
Madam Speaker! I am happy to answer the last one. We received the proposal on short notice, and we did not judge that we, as a transitional government, would be able to bind Sweden to such an agreement. I saw no major complications regarding this, but there were a number of issues that required proper anchoring, and that time did not exist. That time does, however, exist moving forward.
Tomas Brandberg (SD)
Madam Speaker! Thank you, Prime Minister, for the report!
I want to emphasize that the support for Ukraine is broad in this chamber, which I find very pleasing.
Regarding AI, Sweden is, and has long been, one of the world's leading innovation countries. Our technical expertise and the entrepreneurial spirit that characterizes our business life constitute one of the foundations of our prosperity.
There are good conditions to take advantage of the ongoing AI development, but for Sweden and Europe to be able to assert themselves in the global competition, it is required that we provide companies with the right conditions.
Unfortunately, the EU has all too often become a brake rather than an enabler in the digital field. The AI Act and a number of other regulatory frameworks have together created a rather tangled bureaucracy and an uncertainty that risks hindering investments and innovation. This applies in particular to small and medium-sized enterprises.
Madam Speaker! My questions to the Prime Minister are as follows: How is the Swedish government acting concretely in the union to prevent us from hindering our own AI development? How is the government acting to push for more extensive simplification work with fewer overlapping regulatory frameworks, reduced reporting requirements, and greater predictability for companies so that the EU can become an enabler for innovation and competitiveness rather than a hurdle?
Statsministern Ulf Kristersson (M)
Madam Speaker! I would say that we – a very broad we – act in two ways. One is that we act as a role model. Everyone in all of Europe knows that in no other country and in no other place outside of Silicon Valley and possibly also on the Chinese eastern coast is there anything that can compare with the Swedish tech cluster. It is fantastic. This is due, among other things, to good political conditions but above all to the brilliance, talent, and hard work among the entrepreneurs who have built those companies. To be a role model, both technologically and in terms of the capital market, is incredibly important – not least so as not to destroy what is working well.
The second thing is, of course, to ensure that the EU does not do what it unfortunately sometimes excels at, namely regulating things. Swedish input was an important part behind Ursula von der Leyen's simplification agenda and that the EU did not rush ahead with a detailed AI regulation that would make Europe a bad continent for developing AI applications.
I am not saying that we should not have international rules in the future, but that the EU should act completely independently and in practice make Europe a worse place for technology companies would have been a bad idea. In that regard, I think the Commission has been very responsive to Swedish views.
Tomas Brandberg (SD)
Madam Speaker! Thank you, Prime Minister, for the answer!
I want to dwell on the situation of small and medium-sized enterprises. Large companies often have significant resources to monitor and influence the EU's legislative work. They also have administrative and legal resources to handle an increasingly extensive body of rules.
For smaller companies, the situation can be different. There, the same regulatory framework can mean a significantly greater relative burden, both economically and administratively. The risk is therefore that the smaller, innovative companies will find it most difficult to take advantage of the possibilities of AI technology.
My question to the Prime Minister is as follows: How is the government acting to ensure that the perspectives of small and medium-sized enterprises are taken into account, so that regulations are not designed in a way that, in practice, benefits the large, established companies at the expense of smaller and potentially more innovative actors?
Statsministern Ulf Kristersson (M)
Madam Speaker! It is naturally an important question, but I still think that the good ecosystem in Sweden shows that there are good opportunities for smaller actors to act quickly in Sweden. Sweden is a good place for those types of investments and innovations. We shall not settle for that, but it is a good foundation.
The second thing, which is now happening at the European level, concerns what is sometimes called EU Inc. It is a common regulatory framework to, for example, establish companies in Europe as a whole so that smaller companies do not have to apply for permits in 27 different EU countries and can instead enter in one place and resolve many matters at once. Smaller companies have relatively more benefit from this than larger companies. We will have to see how far it goes. There are many countries that are now monitoring their specific national aspects regarding this. There is, therefore, a certain risk of dilution.
I still think, fundamentally, that it is a good idea to have an entry point for new companies that want to establish themselves internationally in all of Europe.
Banesa Martinez Roldán (V)
Madam Speaker! Thank you, Prime Minister, for the report!
During the consultation before the informal meeting, I raised questions about personal privacy, control over one's own data and
the inhabitants' trust in AI development. The Government then emphasized the importance of following and
Evaluate the regulation that already exists. At the same time, we can read that the President of the European Commission speaks of Europe's extensive and high-quality data as a strategic resource to strengthen European AI development and competitiveness. This raises questions about how these ambitions should be balanced against the protection of individuals' data and the control over one's own data.
How does the Prime Minister view the issue of European personal data when data or data processing occurs in third countries, and how do we ensure in this context that individuals continue to have control over their personal data and sensitive data?
Statsministern Ulf Kristersson (M)
Madam Speaker! This is also an important issue. Today, there are a number of European regulations to protect privacy, and there are somewhat different opinions on how well they function. There are also somewhat different user perceptions of how practical they are, but they exist nonetheless.
A problem that is mentioned repeatedly is that a very high proportion of our data is handled and processed in other places. This is one of many reasons why there is talk about building more independent European capacity. In that case, we are talking about having companies that can handle those types of tasks. It is also about building data centers that are large and powerful enough to be able to process such large amounts of data. Then we come to other trade-offs that must be made, for example, those concerning nature, electricity use, and job opportunities.
There are no simple answers here, but one must weigh things against each other.
I would say that access to high-quality data is an underrated thing. Sweden is a small country, but we in Sweden certainly have a handle on our data. We cannot measure ourselves against China when it comes to the amount of data produced every day, but we have historical data. We have a handle on our data, and we have high quality in our data. I believe that this is an underrated asset in a future Europe.
I am otherwise a bit afraid that large countries – China is the most obvious example – which daily produce an extreme amount of data, will gain a head start in everything that AI processes. This is something to keep in mind. That is also a reason to make joint European efforts – the EU has, after all, 450 million inhabitants.
For the sake of form, I should say that none of this received any solution at the meeting. These were things that were ventilated at the meeting regarding AI.
Banesa Martinez Roldán (V)
Madam Speaker! It is of course important to have competitiveness against the USA and China. It is also important to create own capacity and ensure that there is a European sovereignty. As data is used more strategically for AI development, however, questions arise about where data is stored, processed, and used. How does the Prime Minister view the trade-offs when data processing takes place in a third country?
Statsministern Ulf Kristersson (M)
Madam Speaker! Right now we are, as I said, to a very large extent dependent on the USA for this. The very largest providers and owners of cloud services are American. This is nothing that one replaces quickly. Everyone is probably in agreement that it would be very good to have more European alternatives, and I can promise that France with some force put forward that position, including sovereignty or strategic autonomy. This, however, cannot be ordered.
It is very good that a number of large companies are now European. Mistral is one example, and there are also several well-known Swedish companies. No one would rather than they would want to see more competing locations, including in Europe. We must, however, ensure that the competitiveness, capital procurement, and investment conditions in Europe are so good that the companies choose to stay here. In that area, we are not yet home, I would say.
Anna Lasses (C)
Madam Speaker! I thank the Prime Minister for the report.
I will start with AI. I once heard a researcher on the radio talking about the advantage of European bureaucracy. It slows down the speed of AI. Unfortunately, the same bureaucracy does not exist in the USA and China, which means competitive disadvantages for Europe. It is still a thought-provoking idea that bureaucracy can help us humans and make it so that this does not accelerate and completely slip out of our hands.
As the Prime Minister was in, it is about finding a balance, for example when it comes to small businesses. It is a matter of having a stable bureaucracy but not locking the technology itself with various regulations. I wonder if this was something that was taken up.
I also wonder if anything was said about privacy, cultural values, and that whole discussion.
I move on to Ukraine. Is the Prime Minister, after these discussions, confident in the EU's guarantees for support ahead of the winter?
In the morning, I had the privilege of meeting Jevgenij Smirnov, who is a Russian lawyer and human rights activist. His contribution concerned the possibility for Russian refugees to obtain asylum and how this is secured, given that we at the same time do not want to let in Russian infiltrators.
Statsministern Ulf Kristersson (M)
Madam Speaker! There were two important issues.
I will start with AI. It is quite true that an efficient, predictable, and non-intrusive bureaucracy is an advantage in all countries. This should not be underestimated, and here I believe Europe stands quite well. A slow bureaucracy is, however, not always a competitive advantage. We do not want people to shoot from the hip, but it cannot be that everything takes much longer – then many companies will turn to other places.
This is happening right now at a frantic pace – possibly too fast – in the USA and China. They are engaged in a struggle against each other, moreover on completely different terms. One country is still a democracy with traditional companies. The other is very much a state-capitalist country which also has completely different possibilities to regulate the use. I am a bit afraid that Europe will end up in an intermediate position and not be perceived as particularly important in the competition. I believe we should guard against that. There are many fine smaller AI companies in Europe, not least those that make applications, and we shall value their contribution to this.
I move on to the question regarding Ukraine. No, I am not at all confident in this – not in the least. I mentioned the funding earlier. This was not a decision-making meeting, so nothing was decided at the meeting. It is, however, quite obvious that Ukraine needs around 75 billion euro for 2027, of which 45 billion euro is what the EU Commission would like to contribute in a joint loan. Now, a rather heated discussion is taking place. A number of countries want to opt out, as it is called. They say: Be my guest – feel free to fund this, but we do not intend to participate and contribute.
Together with Germany, we are marking very strongly against these countries. At the same time, we are very keen on getting through the support that Ukraine needs. This also enters into the next long-term budget, which is complicated both for Ukraine-related reasons and for completely other reasons.
I believe this will be Ukraine's worst winter during these five years. They must get through it.
Anna Lasses (C)
Madam Speaker! I completely agree with the Prime Minister – we simply must solve this.
I am making a radio transition to MFF. Ukraine must be secured in that as well.
As I understood it, you were talking about the MFF in the corridor. How is it going regarding the frameworks that we discussed earlier?
I will not mince words in saying that the Centre Party feels concern regarding the aid. Ireland has proposed a cut of Global Europe by 17 percent; I do not know if this was mentioned. The Prime Minister is welcome to comment on the discussions regarding the MFF.
Statsministern Ulf Kristersson (M)
Madam Speaker! I would be very happy to do so. When it comes to the MFF, that is, the long-term budget, the situation is totally locked. No changes are happening at all. It reminds me a bit of the Swedish government formation. It is, as said, totally locked, and one gets nowhere. We are a large number of countries that want a more modern budget, substantially lower frameworks and lower costs, not least for those of us who are net contributors. A number of other countries think exactly the opposite and want a budget as traditional as possible that maximizes their benefits. It is, therefore, two different worlds. Furthermore, we have the Ukraine financing in the middle of this, and there the commitment is somewhat varied.
The President of the European Council intends to convene us for a separate non-decision-making meeting regarding only the MFF. I believe it is at the end of next week or in any case soon. He hopes for a decision in December, but many do not believe in a decision in December.
Rebecka Le Moine (MP)
Madam Speaker! I thank the Prime Minister for the report.
Before I move on to AI, I want to thank the government for reflecting the unity and the great support the Riksdag has regarding Ukraine. Regarding the MFF, we will return to it several times, as well as the Miljöpartiets views on it.
AI is often described as a race, but I am a bit afraid that it will become a race to the bottom if we do not include other important values. It is no coincidence that Finland and Sweden are attractive locations for data centers. What is attractive includes, among other things, our shared values, such as the electricity capacity, the landscapes, nature, and the water. We cannot sell this out to the highest bidder, because it comes with a cost that must be paid locally, while the profits, the high profit, are in the global tech giants' calculations. Therefore, it is important that we are cautious and include all these parts in the consequence and risk analyses.
Several countries in the EU and Europe have taken different measures, but I see it as natural that the EU steps forward and acts together in a strategy that secures both competitiveness and the positives of AI as well as the most basic things, such as our security, access to clean water, and access to cheap electricity for consumers.
How did the consensus go regarding gathering together on these large, overarching issues?
Statsministern Ulf Kristersson (M)
Madam Speaker! This was a seminar – and there is nothing wrong with that. On one hand, the Mistral chief described in a very simplified manner all the possibilities that exist with AI, including security and stopping disinformation, and he wants Europe to be a good place for AI. This is, after all, part of his job description. On the other hand, Professor Hinton described in all respects the risks of an uncontrollable AI development that acts beyond human control, which we have seen several examples of recently. Both perspectives are completely legitimate, and it is up to each individual to decide what they believe in most.
We also entered into the constant trade-offs and priorities, and Svenska kraftnät is now, as said, making such an analysis for the Swedish side. Other countries are doing it for their part. I completely share the view that Sweden and Finland are attractive, which is good, but Sweden and Finland should rule themselves. Others' needs should not govern everything. Within the framework of European competition neutrality, we shall decide how Swedish energy and Swedish natural resources best come to the benefit of the Swedes. There we can certainly learn a lot from other countries.
EU strategies can exist in some areas, without doubt in terms of security. There can also be strategies for common computing capacity. The gigafactories that are now being built under the EU's auspices are examples of cooperation, and Sweden and Finland have a common stake in such a proposal.
There are things that can be European competence while other things are national competence, and we probably want to keep those.
Rebecka Le Moine (MP)
Madam Speaker! I thank the Prime Minister for the answer.
One should probably not underestimate a high-level seminar on this subject because AI affects all policy areas.
The Prime Minister's previous answer today as to why Sweden does not stand behind the appeal that Finland and Norway took the initiative for, I think concerns the fact that we have a transitional government. But now that one has a bit more time, can one not, based on the new state of knowledge and the new concerns, try to anchor this policy and direction in time? I believe, in fact, that there is a common concern in Sweden's Riksdag.
Statsministern Ulf Kristersson (M)
Madam Speaker! I take this point. However, it came up very suddenly in the UN General Assembly, and we did not want to immediately join Sweden to an outcry that could cause some discussion in Sweden. I believe it is quite uncontentious to say that international regulatory frameworks are required, even if it is difficult, but that they should automatically fall under the UN is probably not as obvious, especially not if other major actors do not choose that path.
There were, however, discussions, so it was not a matter of lack of interest but of lack of time.
The Nordic and Baltic countries are distinguished by their great interest in the issue, both regarding the opportunities and the risks. Greece and France also show great interest, for various reasons. It could be that the leaders are extra interested and that a number of companies are located there. But I will be honest and say, without pointing at anyone, that there are quite a few countries that are not thinking about this significantly at all.
Erik Ottoson (M)
Madam Speaker! I thank the Prime Minister for the report.
Support for Ukraine is more important than ever, and it is pleasing that the EU is still able to hold together in large parts. In the long term, there are of course challenges within the EU regarding how we can maintain the support, but I feel proud to be Swedish when Ukraine comes up.
So, regarding artificial intelligence. I experience that the discussion often circles around the balancing act between, on one hand, regulating, and on the other hand, being tech-friendly and taking advantage of the benefits of technological development, the influence over it, and the use of it.
It is easy to agree on certain things, for example that a regulatory framework is needed and preferably an international regulatory framework. It may be somewhat more difficult to agree on what the international regulatory framework should aim for. We can be completely certain that the USA and China will have their own views on what the regulatory framework should aim for – likely to favor their own development and position on the market.
What risks becoming even more difficult is landing on what the EU wants the regulatory framework to aim for.
It is an old European habit that one can agree that a regulatory framework is needed, but then the process takes on a life of its own without a clear objective. Therefore, one also cannot evaluate whether it became a good regulatory framework or if it just became a regulatory framework. How did the discussions at the meeting go regarding the purpose of an international regulatory framework and what one wants to achieve with it? How is Europe to assert itself in that discussion?
Statsministern Ulf Kristersson (M)
Madam Speaker! We probably think a bit differently in different European countries, and also somewhat depending on what relative position our countries have in technology issues. I am not ashamed for a moment that I absolutely do not want to see a regulatory framework in Europe that reduces the international competitiveness of Swedish technology companies.
Every time there is a desire for common regulations on how the capital market functions, we see opportunities but also risks. Just look at the number of stock market listings of Swedish tech companies! It is quite obvious that something works well here, without us being flawless in any way.
One does not need to question the purpose, but there are probably two: to increase the common European impact and to increase security awareness and reduce the risks of serious missteps in AI development. These are two very trivial but honorable purposes.
How does one do it best then? There is always a risk that if the EU only has legislation and regulation as tools, then it is applied everywhere. It is difficult to regulate into competitiveness, and it is difficult to regulate into talent. So there is a small risk, but I think it has decreased now, so that we do not run so fast that we acquire a purely European regulation that makes other countries think that Europe is not going to be a good place at all.
If one knows that one has regulatory power and that the whole world must follow suit, one can do so, but I do not believe that the EU currently has enough regulatory power to ensure that the USA and China automatically adapt to the European regulatory framework. That does not mean that we have to do as they do, but one must have respect for the fact that the competitive situation is quite tough.
We have to proceed a bit cautiously and do what we can to ensure that the regulatory frameworks are as international as possible. What is being raised in the discussions is, for example, to invest in risk minimization and to have some kind of authorization model when new AI models are released, but also to have language models that rely heavily on European data, not just on American or Chinese data. There are quite a few things that can be done.
Erik Ottoson (M)
Madam Speaker! I am in complete agreement regarding this. That very balancing act, I believe, is extremely difficult but very important. It is also the answer to many of the questions that have been asked here today. We will therefore not be able to refrain from being a part of the technological development, but we will also not be able to let it be released completely and entirely.
This is very reminiscent of everything from the invention of the automobile once upon a time, which many viewed with skepticism, to social media, which has also proven to be difficult to regulate, and a number of different new phenomena in our world that are becoming increasingly global, disruptive, and fast.
I would therefore like to ask the Prime Minister if he would say that there is an awareness in the European Union that we will need to act relatively quickly in relation to the technological development compared to what we in the EU may be used to doing, and also sometimes swallow the bitterness and not always get exactly what we want.
Statsministern Ulf Kristersson (M)
Madam Speaker! That is probably the case. Then the basic cultures are somewhat different. Now, we are not going to just praise ourselves – we have our shortcomings – but one thing is that we are normally tech-oriented and see transition as a better method to handle new technology.
Competitiveness in Swedish companies means, for example, that many employees can be retained, instead of opposing the technology and hoping that it disappears. There we have a historical strength. Then, one cannot guarantee that the AI revolution will behave exactly the same as other major technological shifts, but I think we have shown our ability. Even with major technological shifts that we have not created ourselves, we have been able to utilize the applications.
In both traditional industry and modern tech companies, we have been good at constantly adapting. Sometimes this is the small exporting country's strategy. We cannot rule the world, but we can adapt to the world, to our own advantage. That is where I believe our focus should lie.
Magnus Berntsson (KD)
Madam Speaker! Thank you, Prime Minister, for the account of the meeting!
It is very appropriate to discuss AI. I believe we all feel that. It has become many aspects that are important for all of us. Geopolitically, it is a power issue. It would be interesting to hear if the Prime Minister can mention something about how Europe is now viewed from the aspect that we have China on one side and the USA on the other. I also want to hear if there was some kind of consensus in the Council on how one views China's commitments in this context.
It is democratic values and Europe's competitiveness that are at stake. It is extremely important that we include both parts, but that we do not, as the EU has had a tendency to do historically, end up in a situation where we regulate away opportunities instead of deviations. It would be interesting to hear how that was discussed.
There are many more areas that I assume were discussed, such as defense, welfare and not least healthcare and the situation for children and young people, but I want to say a few words about Ukraine. We have heard that an extra meeting will be held with the heads of state and government regarding financial issues. Will it be solely about the MFF, or will the situation in Ukraine be addressed a bit more broadly already now? I am thinking of the frozen assets, financial support to Ukraine in a broad sense, budget support and other matters.
Statsministern Ulf Kristersson (M)
Madam Speaker! I shall take it in turn and order – China first. It was, therefore, not a meeting about China, even though we did actually have a meeting about China the time before, when we talked a lot about the trade relationship with China and the new European tools. These will probably not pass unnoticed in relation to China. It is a somewhat more blunt view of trade with China. Then, of course, services are also an important matter.
It is clear that AI is special. A country that is ruled authoritatively, which is technologically successful and has 1.3 billion people, provides unique conditions if one wants to be successful in, for example, artificial intelligence, and the public scrutiny of regulatory frameworks and such is very limited. No one underestimates that, I would say, but we have not yet seen the consequences of it. Therefore, the collective strength in Europe, which has at least half a billion people, must be used.
When it comes to Ukraine, the meeting will be about the long-term budget and the financing for 2027. I am concerned about the trends we are seeing right now. Four countries have stated quite explicitly that they, as it is called, do not want to stand in the way of money but that they want to use their right to opt out. There is no general right to do that, but it was a unique solution so that Hungary would not block the unanimity.
This is not freely chosen work. If one wants to exercise a veto, one can always do so, but one cannot generally say that others should participate in financing and that one oneself should be a grant recipient. Germany was crystal clear on that point, and we others were equally crystal clear. Germany even wrote a letter about it in advance so that no one could misunderstand that position.
Then we had the entire representation. It was good that it was resolved with that, albeit informally. We made no formal decisions, but we now know who will speak jointly for Europe so that we avoid the risk of very many countries wanting to convey just their own point of view to Putin.
Magnus Berntsson (KD)
Madam Speaker! I thank you for the answer.
It would have been interesting to hear a few words about EPC from the perspective that we have been able to show strength and unity from the European side, but that we also have our challenges. How were the discussions with the EPC countries conducted? The expansion and how this will affect the unity moving forward is an extremely important aspect, and it is also the discussion that I understand was held regarding European security in a broad sense.
There is much to discuss. We have experienced what can happen with underwater cables, power supply, and telecommunications. There are many aspects regarding European security, and we perhaps have not quite felt that all countries in Europe's immediate vicinity – potential member countries or candidate countries – have really dealt with them, so far. Were those questions on the table?
Statsministern Ulf Kristersson (M)
Madam Speaker! I arrived a bit late to EPC, for reasons related to the formation of the Swedish government, so I was not there all day. But I was present at one of the more important meetings, the one we had with President Zelenskyy, the new British Prime Minister Andy Burnham, and Canada's Mark Carney, where we went through, above all, the Abu Dhabi talks. So it was not so much about the expansion issues. They are currently, in all respects, about Ukraine, about Moldova, and about Montenegro. That is where we have the expansion, so it was nothing broader.
Then everyone notes that there is a new British interest, which I believe everyone is quite enthusiastic about and perhaps a bit astonished by. But we are careful to say, not least from the Swedish side, that we are for all British ambitions to reconnect with the EU, however complicated they may be.
The meeting with Zelenskyy was fundamentally good, even though we think there are no clear advances anywhere. Then, different leaders are somewhat differently engaged in looking for small glimmers of hope. We are a bit more blunt in Sweden.
Louise Minnhagen (L)
Madam Speaker! Thank you, Prime Minister, for the report!
The Prime Minister mentioned that you in Dublin raised children's rights in relation to AI. This significantly affects the school system because students today use AI tools in their schoolwork.
Earlier in November, the EU Commission presented its education and skills package. One of the proposals is to strengthen teachers' knowledge of AI both in teacher education and in continuing education. The Commission also proposes that AI should be used when it is pedagogically justified and in a way that suits the students' age.
Today, on International Students' Day, I want to particularly highlight the importance of our children first receiving basic knowledge in reading, writing, and arithmetic. But an AI service can help a student who has gotten stuck. It is indeed a fantastic aid. However, the teacher needs to be able to determine when the tool actually helps and what the student has learned. Here, teachers can receive concrete support by sharing experiences with colleagues in other European countries.
My question to the Prime Minister is: Did you discuss at the meeting the importance of the EU working together to implement AI in schools in such a way that teachers receive support and help in their educational work?
Statsministern Ulf Kristersson (M)
Madam Speaker! We did not do it so concretely or in such detail. We had no separate education discussion.
I want to remember that Arthur Mensch described all the educational opportunities with AI, including the need for students to be proficient in this. One can imagine that they skip a number of technological steps so that they do not learn a number of basic things but only rely on AI. Everyone who has ever used a model has surely discovered that the more you can do yourself, the easier it is to see the flaws that exist in AI models. It is easy to let oneself be deceived.
On the risk side, there is naturally also the risk of deepfakes and the risk that children and young people get a very distorted view of reality regarding various things. There are a lot of different risks in this. Some are linked to ordinary social media, where AI-driven algorithms cause one to get a very skewed picture. These risks are easy to see, but we had no exact educational discussion about this.
Louise Minnhagen (L)
Madam Speaker! Thank you, Prime Minister, for the answer!
AI offers fantastic opportunities, but as has been said, there are also some risks. I would like to take the opportunity to bring up another aspect. A student can, for example, start with a math problem and ask AI for a solution. A few minutes later, the student might tell the chatbot something very personal. In that case, it is important how it is built and what happens with the data.
My question is: Did you discuss the importance of negotiating with the AI providers in as adequate a way as possible regarding their responsibility linked specifically to the school environment? You perhaps did not, because you were not so specific, but it is still a very important question.
Statsministern Ulf Kristersson (M)
Madam Speaker! No, we did not do that.
Fundamentally, this is no self-evident EU competence, I must say. Then, of course, one can learn from each other. In many countries, I also believe that different schools have different policies for this and for which tools one allows at all. I believe it is quite a lot up to local decision-making. Then, that does not prevent one from learning from each other when trying to establish which policy one should have.
It was more about how we ensure that Europe is not entirely dependent on American and Chinese technology and American and Chinese data. That was where we were most [focused].
Gurgin Bakircioglu (MP)
Madam Speaker! Thank you to the Prime Minister for the report back!
Not so long ago, TV4 tested Google's AI with the search "I am alone with a German." The answer was reassuring: Germans are not dangerous. They smile and are friendly.
I tested it myself directly afterwards – I wrote "I am alone with a Swede" and received the same friendly tone back, the same as with the German. But something happened when I wrote "I am alone with a Somalian." Then I received the recommendation to seek protection and call 112 if I feel threatened.
The Prime Minister previously mentioned how AI can contribute to the view of humanity. It could also be that we, with our view of humanity, have trained AI through majority norms. TV4 received similar answers when they tested "I am alone with an Arab." The AI wrote: If you feel threatened or are in danger, call 112.
This is not some strange thing on the internet. The internet is a reflection of ourselves. It is we humans who have filled it with the opinions of the majority society and even our most hideous prejudices. When AI is trained on our material, it can also learn to repeat and reproduce these prejudices.
Swedish research has repeatedly shown that people with foreign-sounding names are screened out despite equal merits. AI is used to sort job applications. The Prime Minister has mentioned the labor market here. Even applicants with backgrounds in Somalia and Iraq have been disadvantaged in Swedish recruitment experiments; there is plenty of data on that.
The Prime Minister himself has stated that he uses AI in his work, so I assume that he knows this well and knows how it can contribute to the most despicable things, which we do not want to see.
In Dublin, the EU leaders discussed AI's impact on the labor market and fundamental rights. The EU has postponed specific control requirements for recruitment AI until December 2027.
My question to the Prime Minister is what concrete initiative Sweden promoted in Dublin or intends to promote within the EU so that AI systems that sort job applications are audited for discrimination already now, so that we do not let society's old prejudices decide who gets the job and who does not get the job.
Statsministern Ulf Kristersson (M)
Madam Speaker! This is interesting. To state the obvious: AI can be used to protect our societies but also to harm our societies. It can be used for purposes that are in every sense honorable and for purposes that are in every respect terrible, no doubt about it. In that sense, AI is a tool that reflects the intention of the one who uses it.
But there is a deeper aspect to this. Since AI is built on massive datasets, naturally, values, language, culture, and history are significant when training models. This is one reason why we now, together with both public and private actors, and very many media companies, want to invest in a Swedish language model so that Swedish culture, history, and values are included.
Then we will not be able to influence other models. I can guarantee that the Chinese Deep Seek does not reflect the same values that may exist in the French Le Chat, for example.
I think the member captures an important issue. We were not looking into the labor market-specific European regulations. As far as I know, there are no regulations for this at a European level yet, but it is national legislation that applies for, for example, labor market issues. But I think that what is raised here about the risk of using AI and in that way reinforcing existing biases is undoubtedly there.
Gurgin Bakircioglu (MP)
Madam Speaker! Thank you, Prime Minister, for the answer!
Discrimination is already prohibited in Sweden here and now. Research shows, as I mentioned, that one is screened out from jobs because of their name. We can use other language models. But there is a global trend that one is discriminated against as a minority in the labor market, so even if we were to train a Swedish language model based on the labor market, it would have looked the same. There is research that shows how, for example, a person with an Arabic-sounding name is screened out in favor of someone named Sören, Sixten, or Johanna.
I therefore pose this question: Can the Prime Minister mention a concrete measure that Sweden intends to take to control discrimination in AI systems before the EU's postponed rules begin to apply?
Statsministern Ulf Kristersson (M)
Madam Speaker! No, I cannot do that here and now. I believe you heard in my presentation that there was a rather high level of abstraction when it comes to risks and opportunities with AI. In the same way as in other areas, one can imagine using AI to detect biases, and one can also use AI to reinforce biases. I believe this reflects the description of risks and opportunities quite well.
I have read a very interesting article in Foreign Affairs by James Manyika, who is the research head at Google, which deals with the combination of AI and quantum computers. There, the risks are described that by using the combination of AI and quantum, one could crack all the world's cryptos. Imagine that! Realize the scope of this! And what does James Manyika say in the article? Yes, that it is with AI and quantum that we can also protect the cryptos.
So it is all the time about the balancing act between bad ambitions and protecting important values.
Håkan Svenneling (V)
Madam Speaker! Thank you, Prime Minister, for an interesting presentation!
There has been a lot of AI in the discussion now, but I want to return to the question of Ukraine, which some speakers have also raised. Now Europe and the EU have decided. It is not a question of who should represent Europe but about how many. It has been decided that it should not just be one but a total of seven representatives. These are partly the E5 countries, partly the EU and partly Nato.
My reflection in this situation is: What mandate do these seven have to act externally? And how are these seven different individuals supposed to agree among themselves? What information can we expect them to share with states like Sweden regarding how the negotiations affect us and how they progress? If the individuals do not agree – how is that resolved? It is not all that difficult to imagine that it will be so, as Italy's and Poland's lines when it comes to Ukraine have until now been somewhat different, to put it diplomatically.
I also want to ask the question: Why did the Nordic countries not get a place at the table? The Nordic countries, together with the Baltic countries, are states that provide a large percentage of support but also a large support in real terms to Ukraine. Would it not have been reasonable that we also received a clearer place at the table when other states, which give quite little in comparison to us, are now allowed to be included and follow the peace negotiations more closely?
Statsministern Ulf Kristersson (M)
Madam Speaker! As I believe I said: This is not a perfect solution, but we can live with it. There is no perfect solution.
The first is the obvious: There are a number of pure EU issues, which really are only the EU's and which, among my colleagues, also require consensus. In those cases, it is only the President of the European Council who can speak for the entire EU after an EU decision. This applies to everything involving frozen assets, everything that includes membership issues for, for example, Ukraine and funding. These are EU issues, and it is only the President of the European Council who can ultimately speak for them.
Then there are a number of things where the largest countries must, in practice, agree if we are to have a chance of achieving unity. France and Germany will never outsource to anyone else to speak for them in major geopolitical issues. One may have views on that, but that is how it is.
It is extremely important that some larger countries that are also strong contributors to Ukraine and neighboring countries, namely Poland, are included. Including Poland without including Italy would not have been possible. Keeping Italy in the broad consensus is strategically important. To this comes the United Kingdom. The United Kingdom is not the only European non-EU country. Ask Norway!
As you hear, there is no obvious solution. E5 was probably the best we could get in a situation where more countries felt a need to formalize these things to some extent. I do not distrust for a moment Jonas Gahr Støre's and Alexander Stubb's conversation with Putin, but I am seriously afraid that many want to convey just their own position to Putin in a cacophony of voices. It would only benefit Russia.
Add to that Coalition of the Willing, where every country, including Sweden, nationally decides what it wants to contribute to a possible security guarantee.
Håkan Svenneling (V)
Madam Speaker! It is good for the Swedish Riksdag to hear the Prime Minister's reasoning on this. I myself have for quite a long time been an advocate for the EU to appoint an envoy. It has met some resistance, and this may be the best possible compromise we can get. Several states important for Sweden are indeed represented at the table. I believe it is important that Sweden has clear and direct channels to the representatives and that one presses for Sweden's interests.
I would like the Prime Minister to return to the upcoming extraordinary meeting regarding the support for Ukraine. What measures does Sweden intend to take so that there will be a large support for Ukraine in 2027? We know that the 90-billion loan that was previously granted led to much bilateral support disappearing from other states.
In this, the important central question of the Russian frozen assets, which are so much more money than the 90 or 75 billion that are currently being discussed, comes up. This could make a concrete difference for Ukraine, and at the same time, it would be Russia that would have to pay.
Statsministern Ulf Kristersson (M)
Madam Speaker! Let me first strongly agree that our direct contacts with these countries at all times are very important. I dare say that we are in constant contact with all E5 countries, with the President of the European Council and with Mark Rutte.
I feel very strongly that Sweden is not being neglected in any way. Everyone also knows that Sweden and a few other countries account for a very large relative share of the support; perhaps a slightly too large share.
I also completely share the view that when the 90-billion loan – that is, in euros – went through, the bilateral supports practically decreased. It can also sometimes be about a lack of political support in countries, about polarization in certain countries, and about poor economies in a number of countries. It is such things that really affect the support.
Finally, it is about the frozen assets. We raise the issue all the time. The Belgian government is currently a bit in between, but they are quite clear when it comes to the fact that they do not intend to accept that decisions are made without total risk-sharing. The Commission is trying to find ways around this right now, but the money is earmarked for Ukraine sooner or later.
Robert Olesen (S)
Madam Speaker! Thank you, Prime Minister, for an interesting report back!
I return to AI. You discussed AI's impact on the labor market and on Europe's future. The Växjö company Fortnox announced yesterday that they have invested 5 billion in AI and simultaneously laid off 200 employees. We Social Democrats welcome the technological development, but the workers cannot alone pay the price for this transition.
My question to the Prime Minister is: What concrete initiatives does the Prime Minister intend to take, both within the EU and in Sweden, to ensure that AI development leads to more and better jobs, that workers are given opportunities for skills development and transition, and that the gains from the technological development also come to the employees?
Statsministern Ulf Kristersson (M)
Madam Speaker! I share the view that there are many both opportunities and risks with this. In Europe, there is a tendency to primarily see the risks. That is because there is also a constant expectation of legislation regarding all such things.
We have parties that systematically work with the conditions in the Swedish working life, including transition issues and technology issues, and that is where they do the most good. So that is how it has been in Sweden until now. I would very much dislike having European legislation telling the Swedish labor market how it should handle the technological transition. I believe that would be bad for Sweden.
Fundamentally, it is about seeing the opportunities and the risks, as we have done before, and understanding that it will have consequences. But we must primarily ensure that we meet the challenges with our technology-oriented transition capability.
Robert Olesen (S)
Madam Speaker! I have a follow-up question for the Prime Minister: Does the Prime Minister consider that companies also at the European level have a responsibility to ensure that their employees take part in the transition, manage the technological development, and are equipped for the future jobs when one is simultaneously investing in AI that replaces jobs?
Statsministern Ulf Kristersson (M)
Madam Speaker! It is clear that I consider it, but I also have respect for the fact that different countries have different legislation regarding such matters and that they also have very different traditions. Decisions are made in completely different ways. One example is Germany, which has a completely different regulatory framework. There are very many regulatory frameworks, and if we do not want to be subjected to regulatory frameworks that other countries consider to be self-evident, for example, statutory minimum wages, we should also not have strong opinions on what other countries should do.
I think it is completely obvious that a technological transformation will occur that will pressure many European companies, but it is still happening. Look at the Chinese competition in strategic sectors, for example solar cells and electric cars, where there is an enormous transformation pressure right now. I believe that we in Sweden have good conditions to manage that competition, but we do not want to be forced into regulations that suit other European countries.
Source: The Swedish Parliament. The speeches come from the open data of the Riksdag, translated into English by AI, which may contain errors.
Portraits: Photo: The Swedish Parliament. (opens in a new tab)