The EU Commission intends to present its proposal for the new emission standard “Euro 7” earlier than planned. It affects vehicles with internal combustion engines, and as this type of engine will still go out of production by 2035, it will probably be the last pollutant regulation of its kind. Is the production of combustion engines still worthwhile? Markus Grabitz explains why some models could be in short supply.
Artificial intelligence has only recently entered the realm of the possible. However, due to its potential to intervene in sensitive areas of daily life, experts are already working hard to regulate it. The AI Regulation is attempting to do just that. With it, Europe is still ahead of the game compared to others, according to Antonio Krüger, CEO of the German Research Center for AI, in an interview with Corinna Visser.
The electricity price is rising – according to analysts even more sharply than the gas price. The Bundesbank suspects that economic output in Germany will therefore fall in the winter half-year. However, the EU energy regulator ACER warns of another development in winter: If many households turn up electric heaters instead of gas heating on cold days, this could overload the power grid and thus jeopardize the security of electricity supply. You can read more about this in the News.
With that, I wish you a great start to the day.
The Commission now wants to present its proposal for new emission standards for cars, vans, trucks, and buses around six weeks earlier than last announced. Plans for a tougher emissions standard (“Euro 7”) are due on October 12, according to the Secretariat General’s list of upcoming Commission meetings. Initially, the target date was November 30.
With this, the Commission is expected to present its proposal for the last ever regulation of pollutants for the internal combustion engine in cars and vans. The end of the internal combustion engine in cars and vans is scheduled for 2035. Manufacturers and suppliers are taking a keen interest in the details. They want to know how high the Commission will set the bar for pollutant emissions. For the industry, the question is how much money it will have to invest in exhaust gas purification to stick with combustion technology until 2035. Will the manufacturers’ traditional business model still be viable until 2035?
Vice Commission President Frans Timmermans, who is responsible for the Green Deal, initially advocated a drastic tightening of the limits. Industry Commissioner Thierry Breton, who also has the interests of French manufacturers in mind, put on the brakes.
The Commission had commissioned a consortium of scientists from several countries to develop a proposal. At the end of October 2020, the so-called Clove Consortium presented its proposal to the Commission in an internal meeting. When it leaked out a few days later, the response was devastating. The proposal envisaged a radical reduction in the limits for nitrogen oxides and carbon monoxide, and emissions of methane and nitrous oxide were also to be significantly reduced.
Particular criticism was leveled at the fact that, in real-world road tests, vehicles were expected to comply with the new limits from the first kilometer and under full load on hills, as well as when towing heavy trailers. Euro 6 provides for exceptions in these special situations.
The manufacturers saw the proposal as a general attack on their business model. Many industry representatives predicted that if these standards became law, so much would have to be invested in exhaust gas purification that the production of small cars in Europe would no longer be profitable, and the production of premium vehicles would generate significantly less profit.
In the meantime, Timmermans has apparently started to rethink his position. The reason for this is that both co-legislators – the European Parliament and the Council of 27 environment ministers – have now set 2035 as the end date for new vehicles with combustion engines. Now it is no longer necessary to accelerate the transformation toward e-mobility by radically tightening the pollutant limits for internal combustion engines, they say.
The manufacturers had always warned that if high investments were needed to meet the Euro 7 standards, there would no longer be enough money to switch to e-drives. There are indications that they have found a hearing with the Commission about these arguments. Observers expect that the Commission proposal will now suggest limit values feasible with existing technologies.
Mr. Krüger, is it possible to regulate an area as unexplored as AI in a regulation?
In principle, it’s a good idea. Regulation is conducive to investment in certain types of AI. It’s a way to steer the market in the desired direction – we don’t want Chinese conditions.
What do you mean by that?
Social scoring, surveillance, selective exploitation of AI techniques for certain groups – we don’t want any of that.
Is the timing right, or is the EU too early or too late with the AI Regulation?
We are still ahead of the game. China has just started to develop its own ideas about regulation. The US tends to have a hard time with regulation and tends to regulate things through individual lawsuits. It’s better to have a good regulatory framework. That’s why it’s sensible to do it now.
Is this the right approach?
I don’t think it’s bad that the EU works with risk classes. This has proven successful in other fields, such as medical devices. After all, there are always areas that are more sensitive than others. Of course, it depends on the design. Regulation must not stifle innovation. And that is also my first point of criticism: I feel there is a lack of focus on promoting innovative ideas in AI. For example, it’s important to give small innovative companies more legroom.
Why do small businesses need more leeway?
Otherwise, there is a risk that they will be left behind by such a comprehensive set of rules – and that is what it will probably be. Large companies can afford large legal departments, small ones cannot. In the worst case, they will be sued by established companies for violations. It’s fine to say we want to regulate this area. But the first step must be to promote innovation in the EU. That’s very important to me as a researcher. And so far, that’s been too far down the list.
What else do you criticize?
I have big issues with the narrow definitions of AI systems explicitly named here, and defining what is and what is not an AI system.
Legal experts criticize that the definitions are far too vague.
I am not surprised. However, I think it’s better to ask what you want to protect. These are certain fundamental values. And systems that call these fundamental values into question must be regulated. Actually, it’s completely secondary what AI technology is under the hood. What matters is what comes out of it. If, on the other hand, AI systems are defined so precisely, there will end up being a lot of disputes and loopholes because someone has developed a new AI technology, and then it doesn’t fall under the regulation.
Isn’t it enough to have a regularly updated annex for such technical developments?
I would be more in favor of thinking about regulation from the end and defining risk classes, areas of application, and values that you want to protect. Then certain loopholes would not arise. Neurosymbolic AI – i.e., the combination of different AI techniques – does not even appear in the regulations, for example. There will always be companies that want to circumvent regulation, and if they don’t use any of the defined AI systems, they can do whatever they want. At the same time, of course, it is also innovation-inhibiting if you already have scissors in your head that then prevents the use of certain techniques.
Why is it inhibiting innovation to look for new solutions?
Because certain promising technologies are then not developed here but in other parts of the world. I find it very difficult to come up with a clear definition of AI systems. Additionally, AI is developing in an incredibly dynamic way. At the moment, completely new processes are emerging that don’t even fall under the aforementioned description. Just think of quantum computing. The situation is similar when it comes to listing areas of application. In principle, that’s okay, but here, too, the regulator must say how it intends to ensure innovation in these areas and not just prevent it.
What do you suggest?
I would scrap the whole AI systematics and ask much more: What do the systems do? Do they influence the autonomy of the individual? Do they interfere very strongly with personal rights? I think the description of areas of application and the classification into risk areas are good steps, even if some are defined too precisely and others too vaguely. But a compromise can certainly be found here.
For example?
For example, in the area of social scoring. Here, the regulators have really thought from the end: systems that classify people based on their socially compliant behavior, for example, are classified as high-risk. Or systems that make autonomous decisions about critical infrastructures. I can understand that very well. But there I don’t need to waste a sentence on how they’re built. If a critical security system starts rolling dice, I guarantee there’s no AI system behind the decisions, but you still want to regulate that for sure. What I’m saying is that conventional systems can also be used in high-risk areas that may perform worse than AI systems but are not subject to regulation.
Does that mean the most important point is the risk classification?
Yes, I would see it that way. It’s important that it’s done well. I would be much more relaxed about support systems where the human being retains control at the end than about systems that act autonomously.
How should the gradation be done specifically?
I would base this not only on the applications but also on how strong the assistance character of the system is. The more it is involved, the greater the gain in safety and comfort, but also the potential risk. But of course, it also depends on the area of application. That’s why it will probably be a mixture of both areas. For example, there’s this absurd discussion about manipulative techniques. We are all psychological beings who can be influenced. All advertising is manipulative behavior. But does it really make a difference if an AI system is behind it or if a product is pushed on us in some other way? And on the other hand, how significant is the damage? Here, the classification is still too fuzzy for me. Here, too, it depends very much on the area of application.
Do your assessments find a hearing in Brussels?
At the moment, we are very much involved via CLAIRE, the AI laboratory network association of researchers and users, and also via other associations. So, we are already being heard, and I assume that there will still be changes to the regulation.
Antonio Krüger is CEO and scientific director of the German Research Center for Artificial Intelligence (DFKI) and scientific director of the research area “Cognitive Assistance Systems” at DFKI. He is an internationally renowned expert in human-machine interaction and artificial intelligence.
These are challenging times for Mauricio Vargas. In late June, the Greenpeace financial expert still spoke of an encouraging signal. The members of the Environment and Economic Affairs Committee had just voted by a majority against the inclusion of gas and nuclear power in the EU taxonomy, which is intended to stimulate investment in green energies.
Less than two weeks later, it was clear that the signal had no effect. In the end, there were not enough votes of the EU parliamentarians to stop the expansion of the EU taxonomy. Starting next January, investments in certain gas and nuclear power plants in the European Union can be classified as climate-friendly. “This will make it easier for the financial industry to simply relabel existing financial products green,” Vargas criticizes.
For almost two years, the economist has been campaigning at Greenpeace for a greener financial world. Among other things, Vargas criticized climate-damaging Bitcoin mining. “Every Bitcoin user should ask themselves whether Bitcoin is worth risking the stability of our ecosystems,” he warns, for example.
Vargas also lashed out at the European Central Bank. It was too slow on climate protection and favored carbon-heavy companies with its bond purchases. Most recently, the 42-year-old called for a price cap on Russian gas imports. Similar to a punitive tariff, this would make exports to Germany more expensive for Russia. The money could be secured in a fund and used for reconstruction in Ukraine, the environmental activist demanded.
And now the EU taxonomy: The EU has missed the opportunity to create credible minimum standards for financial companies, he complains. The greenwashing allegations against Deutsche Bank subsidiary DWS have just shown where this can lead. Following accusations that the fund company was marketing financial products as too green, police and public prosecutors searched Deutsche Bank and DWS in June of this year.
Vargas knows the financial industry well: He worked for a fund company for ten years after he studied international economics in Tübingen and then earned his doctorate on financial stability and currency crises in developing countries and emerging markets. He spent 1.5 years in Mexico during his studies. A country to which Vargas also has a personal connection – he was born in Mexico City.
The fact that he was able to get to know the perspective of a less prosperous country that had already experienced several severe economic crises left its mark on the economist. In 2020, after ten years, he quit the Frankfurt-based fund company and started working for Greenpeace. “It gnawed at me for a long time that I couldn’t live up to my standards of what I actually wanted to achieve,” he says.
Today, he speaks of a successful “change of sides,” even if not much has changed for him content-wise. He continues to work on economic analyses. The only difference is that they no longer help with investment decisions. Instead, Vargas uses them to support the political demands of the environmental organization. For the step “out of the golden cage,” as he calls it, he gives up half of his previous salary.
One of the most important campaigns he has accompanied at the environmental organization so far was the one on the ECB. Last year, Greenpeace literally laid siege to the European Central Bank (ECB): Sometimes they symbolically melted ice in front of the building, sometimes they landed on the premises of the central bank with paragliders. The primary accusation was that the ECB invested billions in energy companies that were damaging the climate.
Vargas himself is usually not actively involved in such campaigns on the ground. That is done by volunteers, he says. But he briefs the activists and assists with press work – with success. Last summer, the ECB announced a change of course: Starting in October, the central bank plans to spend at least €30 billion a year in a more climate-friendly way.
By now, it is also clear that the economist’s sharp criticism of the EU taxonomy will not stop there: Greenpeace has announced plans to file a lawsuit. Before doing so, the organization will submit a formal request to the Commission for an internal review. As soon as this is completed and leads to a negative result, the environmentalists want to take the matter to the European Court of Justice. Vargas and his colleagues will coordinate the lawsuit. Pauline Schinkels
The EU Commission intends to present its proposal for the new emission standard “Euro 7” earlier than planned. It affects vehicles with internal combustion engines, and as this type of engine will still go out of production by 2035, it will probably be the last pollutant regulation of its kind. Is the production of combustion engines still worthwhile? Markus Grabitz explains why some models could be in short supply.
Artificial intelligence has only recently entered the realm of the possible. However, due to its potential to intervene in sensitive areas of daily life, experts are already working hard to regulate it. The AI Regulation is attempting to do just that. With it, Europe is still ahead of the game compared to others, according to Antonio Krüger, CEO of the German Research Center for AI, in an interview with Corinna Visser.
The electricity price is rising – according to analysts even more sharply than the gas price. The Bundesbank suspects that economic output in Germany will therefore fall in the winter half-year. However, the EU energy regulator ACER warns of another development in winter: If many households turn up electric heaters instead of gas heating on cold days, this could overload the power grid and thus jeopardize the security of electricity supply. You can read more about this in the News.
With that, I wish you a great start to the day.
The Commission now wants to present its proposal for new emission standards for cars, vans, trucks, and buses around six weeks earlier than last announced. Plans for a tougher emissions standard (“Euro 7”) are due on October 12, according to the Secretariat General’s list of upcoming Commission meetings. Initially, the target date was November 30.
With this, the Commission is expected to present its proposal for the last ever regulation of pollutants for the internal combustion engine in cars and vans. The end of the internal combustion engine in cars and vans is scheduled for 2035. Manufacturers and suppliers are taking a keen interest in the details. They want to know how high the Commission will set the bar for pollutant emissions. For the industry, the question is how much money it will have to invest in exhaust gas purification to stick with combustion technology until 2035. Will the manufacturers’ traditional business model still be viable until 2035?
Vice Commission President Frans Timmermans, who is responsible for the Green Deal, initially advocated a drastic tightening of the limits. Industry Commissioner Thierry Breton, who also has the interests of French manufacturers in mind, put on the brakes.
The Commission had commissioned a consortium of scientists from several countries to develop a proposal. At the end of October 2020, the so-called Clove Consortium presented its proposal to the Commission in an internal meeting. When it leaked out a few days later, the response was devastating. The proposal envisaged a radical reduction in the limits for nitrogen oxides and carbon monoxide, and emissions of methane and nitrous oxide were also to be significantly reduced.
Particular criticism was leveled at the fact that, in real-world road tests, vehicles were expected to comply with the new limits from the first kilometer and under full load on hills, as well as when towing heavy trailers. Euro 6 provides for exceptions in these special situations.
The manufacturers saw the proposal as a general attack on their business model. Many industry representatives predicted that if these standards became law, so much would have to be invested in exhaust gas purification that the production of small cars in Europe would no longer be profitable, and the production of premium vehicles would generate significantly less profit.
In the meantime, Timmermans has apparently started to rethink his position. The reason for this is that both co-legislators – the European Parliament and the Council of 27 environment ministers – have now set 2035 as the end date for new vehicles with combustion engines. Now it is no longer necessary to accelerate the transformation toward e-mobility by radically tightening the pollutant limits for internal combustion engines, they say.
The manufacturers had always warned that if high investments were needed to meet the Euro 7 standards, there would no longer be enough money to switch to e-drives. There are indications that they have found a hearing with the Commission about these arguments. Observers expect that the Commission proposal will now suggest limit values feasible with existing technologies.
Mr. Krüger, is it possible to regulate an area as unexplored as AI in a regulation?
In principle, it’s a good idea. Regulation is conducive to investment in certain types of AI. It’s a way to steer the market in the desired direction – we don’t want Chinese conditions.
What do you mean by that?
Social scoring, surveillance, selective exploitation of AI techniques for certain groups – we don’t want any of that.
Is the timing right, or is the EU too early or too late with the AI Regulation?
We are still ahead of the game. China has just started to develop its own ideas about regulation. The US tends to have a hard time with regulation and tends to regulate things through individual lawsuits. It’s better to have a good regulatory framework. That’s why it’s sensible to do it now.
Is this the right approach?
I don’t think it’s bad that the EU works with risk classes. This has proven successful in other fields, such as medical devices. After all, there are always areas that are more sensitive than others. Of course, it depends on the design. Regulation must not stifle innovation. And that is also my first point of criticism: I feel there is a lack of focus on promoting innovative ideas in AI. For example, it’s important to give small innovative companies more legroom.
Why do small businesses need more leeway?
Otherwise, there is a risk that they will be left behind by such a comprehensive set of rules – and that is what it will probably be. Large companies can afford large legal departments, small ones cannot. In the worst case, they will be sued by established companies for violations. It’s fine to say we want to regulate this area. But the first step must be to promote innovation in the EU. That’s very important to me as a researcher. And so far, that’s been too far down the list.
What else do you criticize?
I have big issues with the narrow definitions of AI systems explicitly named here, and defining what is and what is not an AI system.
Legal experts criticize that the definitions are far too vague.
I am not surprised. However, I think it’s better to ask what you want to protect. These are certain fundamental values. And systems that call these fundamental values into question must be regulated. Actually, it’s completely secondary what AI technology is under the hood. What matters is what comes out of it. If, on the other hand, AI systems are defined so precisely, there will end up being a lot of disputes and loopholes because someone has developed a new AI technology, and then it doesn’t fall under the regulation.
Isn’t it enough to have a regularly updated annex for such technical developments?
I would be more in favor of thinking about regulation from the end and defining risk classes, areas of application, and values that you want to protect. Then certain loopholes would not arise. Neurosymbolic AI – i.e., the combination of different AI techniques – does not even appear in the regulations, for example. There will always be companies that want to circumvent regulation, and if they don’t use any of the defined AI systems, they can do whatever they want. At the same time, of course, it is also innovation-inhibiting if you already have scissors in your head that then prevents the use of certain techniques.
Why is it inhibiting innovation to look for new solutions?
Because certain promising technologies are then not developed here but in other parts of the world. I find it very difficult to come up with a clear definition of AI systems. Additionally, AI is developing in an incredibly dynamic way. At the moment, completely new processes are emerging that don’t even fall under the aforementioned description. Just think of quantum computing. The situation is similar when it comes to listing areas of application. In principle, that’s okay, but here, too, the regulator must say how it intends to ensure innovation in these areas and not just prevent it.
What do you suggest?
I would scrap the whole AI systematics and ask much more: What do the systems do? Do they influence the autonomy of the individual? Do they interfere very strongly with personal rights? I think the description of areas of application and the classification into risk areas are good steps, even if some are defined too precisely and others too vaguely. But a compromise can certainly be found here.
For example?
For example, in the area of social scoring. Here, the regulators have really thought from the end: systems that classify people based on their socially compliant behavior, for example, are classified as high-risk. Or systems that make autonomous decisions about critical infrastructures. I can understand that very well. But there I don’t need to waste a sentence on how they’re built. If a critical security system starts rolling dice, I guarantee there’s no AI system behind the decisions, but you still want to regulate that for sure. What I’m saying is that conventional systems can also be used in high-risk areas that may perform worse than AI systems but are not subject to regulation.
Does that mean the most important point is the risk classification?
Yes, I would see it that way. It’s important that it’s done well. I would be much more relaxed about support systems where the human being retains control at the end than about systems that act autonomously.
How should the gradation be done specifically?
I would base this not only on the applications but also on how strong the assistance character of the system is. The more it is involved, the greater the gain in safety and comfort, but also the potential risk. But of course, it also depends on the area of application. That’s why it will probably be a mixture of both areas. For example, there’s this absurd discussion about manipulative techniques. We are all psychological beings who can be influenced. All advertising is manipulative behavior. But does it really make a difference if an AI system is behind it or if a product is pushed on us in some other way? And on the other hand, how significant is the damage? Here, the classification is still too fuzzy for me. Here, too, it depends very much on the area of application.
Do your assessments find a hearing in Brussels?
At the moment, we are very much involved via CLAIRE, the AI laboratory network association of researchers and users, and also via other associations. So, we are already being heard, and I assume that there will still be changes to the regulation.
Antonio Krüger is CEO and scientific director of the German Research Center for Artificial Intelligence (DFKI) and scientific director of the research area “Cognitive Assistance Systems” at DFKI. He is an internationally renowned expert in human-machine interaction and artificial intelligence.
These are challenging times for Mauricio Vargas. In late June, the Greenpeace financial expert still spoke of an encouraging signal. The members of the Environment and Economic Affairs Committee had just voted by a majority against the inclusion of gas and nuclear power in the EU taxonomy, which is intended to stimulate investment in green energies.
Less than two weeks later, it was clear that the signal had no effect. In the end, there were not enough votes of the EU parliamentarians to stop the expansion of the EU taxonomy. Starting next January, investments in certain gas and nuclear power plants in the European Union can be classified as climate-friendly. “This will make it easier for the financial industry to simply relabel existing financial products green,” Vargas criticizes.
For almost two years, the economist has been campaigning at Greenpeace for a greener financial world. Among other things, Vargas criticized climate-damaging Bitcoin mining. “Every Bitcoin user should ask themselves whether Bitcoin is worth risking the stability of our ecosystems,” he warns, for example.
Vargas also lashed out at the European Central Bank. It was too slow on climate protection and favored carbon-heavy companies with its bond purchases. Most recently, the 42-year-old called for a price cap on Russian gas imports. Similar to a punitive tariff, this would make exports to Germany more expensive for Russia. The money could be secured in a fund and used for reconstruction in Ukraine, the environmental activist demanded.
And now the EU taxonomy: The EU has missed the opportunity to create credible minimum standards for financial companies, he complains. The greenwashing allegations against Deutsche Bank subsidiary DWS have just shown where this can lead. Following accusations that the fund company was marketing financial products as too green, police and public prosecutors searched Deutsche Bank and DWS in June of this year.
Vargas knows the financial industry well: He worked for a fund company for ten years after he studied international economics in Tübingen and then earned his doctorate on financial stability and currency crises in developing countries and emerging markets. He spent 1.5 years in Mexico during his studies. A country to which Vargas also has a personal connection – he was born in Mexico City.
The fact that he was able to get to know the perspective of a less prosperous country that had already experienced several severe economic crises left its mark on the economist. In 2020, after ten years, he quit the Frankfurt-based fund company and started working for Greenpeace. “It gnawed at me for a long time that I couldn’t live up to my standards of what I actually wanted to achieve,” he says.
Today, he speaks of a successful “change of sides,” even if not much has changed for him content-wise. He continues to work on economic analyses. The only difference is that they no longer help with investment decisions. Instead, Vargas uses them to support the political demands of the environmental organization. For the step “out of the golden cage,” as he calls it, he gives up half of his previous salary.
One of the most important campaigns he has accompanied at the environmental organization so far was the one on the ECB. Last year, Greenpeace literally laid siege to the European Central Bank (ECB): Sometimes they symbolically melted ice in front of the building, sometimes they landed on the premises of the central bank with paragliders. The primary accusation was that the ECB invested billions in energy companies that were damaging the climate.
Vargas himself is usually not actively involved in such campaigns on the ground. That is done by volunteers, he says. But he briefs the activists and assists with press work – with success. Last summer, the ECB announced a change of course: Starting in October, the central bank plans to spend at least €30 billion a year in a more climate-friendly way.
By now, it is also clear that the economist’s sharp criticism of the EU taxonomy will not stop there: Greenpeace has announced plans to file a lawsuit. Before doing so, the organization will submit a formal request to the Commission for an internal review. As soon as this is completed and leads to a negative result, the environmentalists want to take the matter to the European Court of Justice. Vargas and his colleagues will coordinate the lawsuit. Pauline Schinkels
