Convinced Europeans and democrats can only be disturbed by what could be heard at the AfD’s European election meeting in Magdeburg over the weekend. Top candidate Maximilian Krah ranted about family albums and ethnicity, while co-party leader Alice Weidel wants to turn Europe into a “fortress.” Several successful candidates in the top positions had been supported by the party’s far-right camp.
Its figurehead Björn Höcke provoked with the statement: “This EU must die, so that the true Europe can live.” Allegedly, he wants a new European confederation. Also not off the table is the demand for Germany to leave the Union. The adoption of the election program could drag on into January; in any case, more candidates are to be nominated this Friday.
The divisive potential is also evident when looking at the AfD’s role models. The party presents itself as Putin’s mouthpiece, said CSU Secretary General Martin Huber. Krah is criticized for his closeness to China, and several candidates cited the right-wing populist Orban as a role model.
In the coming months, it will also be important to expose the language AfD politicians use when they talk about “globalists” or “immigration.” False statements are also likely to increase between now and the election date, as was already evident in Magdeburg: whether on the imagined “abolition of cash” or on misrepresented quotas from the asylum compromise. What is needed are thus alert eyes and ears and a clear European course.
All over the world, governments are grappling with how to harness the benefits of artificial intelligence while minimizing the risks of this powerful technology. Germany is relying on the EU, which is currently negotiating the AI Act. Once in force, this will also become directly applicable law in Germany. “We are not planning any national AI regulation in Germany,” says Anna Christmann (Green Party), the Federal Ministry of Economics’ Commissioner for the digital economy and startups, in an interview with Table.Media.
Volker Wissing’s (FDP) digital ministry also sees no reason for Germany to go its own way. “We want to back developers of AI in our country by simplifying access to data and creating clarity about future standards,” says a BMDV spokesperson. “Purely national rules are counterproductive here.”
The German government is thus involved in various international regulatory processes – with distributed responsibilities. The Digital Ministry is in charge of the G7 Hiroshima Process on generative AI, while the Economics and Justice Ministries are leading the negotiations on the AI Act in the EU. As an export nation, Germany is committed to international standards wherever possible “so that developments from Germany are compatible and vice versa,” says the BMDV spokesperson. There, it is important not to slow down developments “by imposing too many complicated rules. That is why Germany is campaigning at the G7 level “for clear transparency rules that leave room for innovation.”
Christmann also says: “We must not stifle development at this early stage. Otherwise, there is a risk that Europe will become unattractive for AI developers.” However, he says it is also clear that we need guardrails for AI that affect fundamental rights. “Something like social scoring or warrantless biometric identification in public spaces has no place in a democracy.”
Christmann has just returned from a trip to the US, where she met AI researchers and entrepreneurs, among others, in California. “In the US, there is no understanding of regulating this innovative technology so harshly,” she says. “We need to balance regulation with encouraging research and development.” She says it is also crucial to keep in mind that AI companies in Europe are just emerging, so Europe has different conditions than, say, the United States. Nevertheless, “We must not always look enviously at the USA. Our startup scene is improving.”
Christmann sees it ambitious that the Americans have not waited until the G7 or the EU have agreed on guard rails or rules for AI but have instead brought their companies to a voluntary commitment. “It’s important to be in exchange with the big technology companies, precisely to avoid bias.” And it is right to put the already existing models on a reasonable footing, she said. “But I don’t share the opinion that a voluntary commitment alone is sufficient,” she stresses.
After the summer break in Brussels, the trilogue negotiations enter the hot phase. The aim is to conclude them under the Spanish Council presidency. Christmann also believes it is important to conclude the negotiations on the AI Act promptly to have reliable regulation. “As long as it is unclear what is coming, investors are holding back. That’s not a good situation.” The law must be ready before the European elections, she said.
Germany had made some comments on the general direction of the Council, including that only relevant and proportionate requirements should apply to general purpose AI (GPAI) systems. And pointed out that the review was not yet completed. In fact, this was before the discussion on ChatGPT reached the public.
That means the discussions – including in the Council – are not over. “We have to ask ourselves whether regulation is fulfilling its purpose of shaping technology for the benefit of people or whether it is stifling development,” Christmann cautions. It is good to limit regulation to high-risk applications, she says. At the moment, however, she says, Europe is all about increasing detailed regulation. “That leads to more difficult conditions, especially for young AI companies like we have in Europe.”
For example: “I cannot see why companies should be required to carry out an additional assessment of the impact of fundamental rights, as proposed by the EU Parliament,” criticizes Christmann. Because that would be time-consuming to impossible, especially for young companies. However, it was the Greens/EFA group that introduced the Fundamental Rights Assessment into the legislative proposal.
What else is important for Germany in the coming negotiations? “To create good development conditions for SMEs and startups,” says Christmann. “We need experimental fields. It must be possible to try things before full regulation takes hold.” That is why, she says, the German side attaches importance to the establishment of real labs and that there are time exceptions for SMEs.
Another point is the creation of exceptions within the high-risk areas. So that, for example, cleaning robots in hospitals are not subject to the same security requirements as patient records, for example. “Here we are working on solutions, the discussion is still ongoing,” Christmann said.
Since the Council had already decided on its General Approach before the discussion on Foundation Models or generative AI (such as ChatGPT) gained momentum, member states had hardly shed any light on the topic. “It is right that the EU Parliament has taken up Foundation Models,” Christmann says. But the German government is still in talks about which regulation is appropriate. The framework conditions have changed much more than usual since the Council position was adopted. Therefore, the need to work on concrete solutions is greater.
Digital Minister Wissing has already made a proposal in the FAZ: He envisions mandatory self-regulation for generative AI. However, he also warns against leaving the formulation to the large US tech companies alone. In addition, there should be a review after just two years. Then it would be possible to assess whether a code of conduct for generative AI is sufficient as a supplement to the AI Act – or whether tough regulation and integration into the law are needed.
Meanwhile, the discussion about the implementation of the AI Act has not yet begun. But Christmann can already say one thing about it: “We want to avoid the mistakes of the GDPR and not have 16 different interpretations of the rules again.” The solution could be a central office, an AI Office – something like the one proposed by the EU Parliament. But there is a need for clarification there, too. “A European AI Office must not become the eye of the needle for AI approval throughout Europe,” Christmann warns.
Professor Uszkoreit, you recently founded a German ChatGPT start-up with a presence in China. What is the objective?
We established the start-up Nyonic in Berlin, with a research team in Shanghai. Our aim is to address two weaknesses of ChatGPT. Current language models are predominantly focused on English and perhaps Chinese. We intend to change that. Additionally, we want to transform ChatGPT into a reliable work tool for specific industries, which it currently isn’t.
But why different languages? English is a global language, and Chinese is heading in that direction.
Much European knowledge is not yet available in English, but only in various European languages. Additionally, AI systems should bring the same benefits to users worldwide. Language diversity is a significant step for ChatGPT and crucial for a multipolar world. The majority of the world is multilingual. The US can afford Anglo-centric technology, and China can afford China-focused technology. However, the majority of the world, including Southeast Asia, Africa, and Europe, relies on covering numerous languages and cultures.
Why do you take on this effort? At 73 years old, you are already at retirement age. You are and remain a pioneer in European AI. You could relax.
Everyone on our team, regardless of age, wants to be part of the biggest AI revolution of our lifetime. Additionally, we wish for Europe not to be left behind by the US and China this time. Europe possesses many intelligent individuals, and numerous innovations have come from here. But then we repeatedly failed. China already has more than 80 language models similar to ChatGPT, while the EU only has three. In Germany, there’s only one, which is not yet among the best. We aim to change that.
The topic of artificial intelligence is not new. You have been working on it for over 30 years. What’s different about ChatGPT that has sparked so much excitement now?
The system has achieved something we call understanding, similar to humans. It’s defined differently for the machine since it lacks our holistic worldview and doesn’t adapt its knowledge based on user input. But the effect is the same: The machine responds as if it truly comprehends the inputs.
What are ChatGPT’s remaining weaknesses?
Although ChatGPT is exceptionally intelligent, it still exhibits surprising knowledge gaps that it occasionally fills with imagination. We cannot fully rely on its knowledge and judgment yet.
Will the next version fix these issues?
You might be surprised to learn that there’s currently no reliable method to train the machine to get exactly what I want. It’s all still trial and error, more of an art than a science. There’s no curriculum; I simply feed data into giant tubes and see what happens. We don’t truly know what the machine is doing in the process. When the results come, we say: Well learned, but dear machine, you shouldn’t believe everything you’ve read. And don’t say everything you believe. Then we specifically feed data to correct or improve the machine’s behavior and see what happens.
AI can be misused by malicious people. That needs to be prevented, right?
Yes, we need rules, especially for the specific critical uses of AI. However, those who design these rules should resist the temptation to regulate such a complex technology as a whole. By the time the law takes effect, AI will have progressed further. Currently, I am more concerned about excessive regulation than evil AI systems. This danger is much greater in Europe than in the US or China.
What is the consequence?
Even research aimed at taming and controlling increasingly powerful AI through appropriate control technologies could be hindered by excessive regulation. The most dangerous are always the meta-regulators who want rules to apply universally.
But Germany has a constitution, a basic law.
It doesn’t make sense to invent a kind of basic law for a new technology that is rapidly evolving when we don’t yet know where it’s heading. It is much more sensible to regulate AI for specific application areas. For example, AI for driving, AI for the pharmaceutical industry, for aviation, or schools. This not only makes sense but is also crucial. Each area has very different application conditions. The Chinese do it this way too. Then, in the distant future, we can distill a kind of AI constitution from the intersection of these individual regulations.
AI will change the world as profoundly as the invention of electricity. If I say we need central regulations for the use of electricity, you immediately realize that’s nonsense. Or the steam engine. If I use it in a factory, it requires different rules than if I use it in a locomotive. AI remains a tool of humans and is not a competitor to humans.
Why is this fear, especially in Europe, particularly in Germany and France, so pronounced compared to Asia and the USA?
That is a good question, and it is essential to acknowledge this fact. Perhaps it is a desperate attempt by Europeans to compensate for their waning influence in the world by trying to dominate the meta-level with their rules. The stronger desire for regulation has grown as they no longer have world powers. In Europe, especially in Germany, some medications require prescriptions, while in the US, they can be bought at the supermarket. In Europe, the focus is more on improbable exceptions than on daily realities. There seems to be almost a great pleasure in worrying about risks rather than bringing technology to success through skill and full commitment.
In the US and China, they pragmatically focus on the most likely scenarios.
In the EU, the division of political tasks between member states and centralized regulation has led to a very particular development. Instead of aiming for a modern, flexible combination of codified law and case law that suits the dynamics and diversity of our continent, this distribution of power has pushed the European continental legal system of codification to the extreme with a mammoth effort.
This tendency combines with the hope that legislation can rid the world of evil and impose one’s own ethics through regulation. This is now holding us back, and the world, led by the BRICS countries, is no longer interested in adopting the rules. Furthermore, fear is being fueled more in Germany than in other countries. In the meantime, a new profession of self-proclaimed AI ethicists, AI lawyers and algorithm hunters has emerged. They make a living by spreading fear. These are people who, strictly speaking, do not know much about AI. I am also unsure if they are the best jurists and social scientists in this field.
The other extreme would not be favorable either: every application needs separate certification because little is centrally regulated.
That is the concern of SMEs and particularly of craftspeople and small businesses. If they have to certify everything separately, it takes so long and becomes so expensive that SMEs say, “It’s not worth it for us.” So, we must limit the certification requirements to the truly critical areas and burden producers and users with the burden of proof only where there are actual exceptions and reasons for concern. And provide cost-effective testing procedures with the help of technology. Above all, we must consider what is practical in everyday life. I am now very curious whether Germany and the EU will get it right this time or act as if they can still determine the rules of the world.
Hans Uszkoreit, 73, has worked in Germany, the US and China as a university professor, research manager, industry advisor, and co-founder of several start-ups. He is considered one of Europe’s leading AI researchers. Uszkoreit is the scientific director at the German Research Center for Artificial Intelligence (DFKI). He initiated and coordinated German and international research collaborations and led several renowned European projects. Uszkoreit has authored over 250 international publications. Recently, he co-founded the Berlin-based ChatGPT-like start-up “nyonic” together with a research team in Shanghai. His wife, Xu Feiyu, is also a prominent AI researcher and was previously the head of AI at SAP.
*The second part of the interview can be found in China.Table.
The EU Commission wants to establish a European satellite constellation. The IRIS2 program – Infrastructure for Resilience, Interconnection and Security by Satellites – stands for securing Europe’s sovereignty. Between 2025 and 2027, up to 170 LEO satellites are to be placed in low Earth orbit. More than €3 billion in public funding is being provided by Brussels and the European Space Agency (ESA).
Only European companies are to be involved in the program. The aim is to be able to act autonomously in the event of a crisis. Thierry Breton, the Internal Market Commissioner responsible for 2022, emphasized: “This is historic! After Galileo and Copernicus, we are adding a third constellation to our European portfolio of strategic space infrastructures! […] IRIS² will be a New Space constellation … integrating the know-how of the major European space industries-but also the dynamism of our start-ups, who will build 30% of the infrastructure!”
As a start-up building innovative dual-use satellites, Reflex Aerospace had high hopes for IRIS². But already during the EU legislative process, concerns arose that the program would primarily serve France’s interests and its well-positioned space corporations. Germany, on the other hand – as the largest contributor to both the EU and ESA – advocated from the outset for the participation of SMEs (small and medium-sized enterprises) and companies from the fast-growing new space ecosystem.
This comprises young companies with fresh ideas that, thanks to their optimized processes and improved products, challenge the traditional players in space. In Europe, these are mostly large (semi-) state-owned companies. In Germany, in particular, a structurally significant new space industry has emerged. By means of micro-launchers, its own satellite construction and laser communication, it is in a position to build its own constellation!
It was explicitly intended to involve SMEs and new space companies in IRIS². At least, that is what the Regulation of the European Parliament and of the Council of the European Union of March 15, 2023 (EU 2023/588) says: “The program [IRIS² should] maximize the exploitation of innovative and disruptive technologies as well as novel business models developed by the European space ecosystem, including New Space, in particular by SMEs, mid-cap companies and start-ups …”. For subcontracts worth more than €10 million, the goal is to “ensure that at least 30 percent of the contract value is subcontracted through competitive bidding at various levels to companies that are not part of the prime bidder’s group…”!
So much for the theory. Unfortunately, “the little ones” were unable to apply directly in the bidding process for the main part of the constellation. The process started on March 23, 2023, and one of the conditions for participation set by the EU Commission was a minimum annual turnover of €500 million. As early as May 2, 2023, Europe’s established space groups – including Airbus Defence and Space, Eutelsat, Hispasat, OHB, SES, Telespazio and Thales Alenia Space – announced the formation of an “open” consortium. This concentration on the long-established space companies – without competition from a second consortium and excluding New Space – is dangerous for Europe’s innovative strength!
Under IRIS², corporations are now able to bypass and keep down unwelcome competitors with appropriate subcontracts. Even subcontracts from new space companies or SMEs remain optional. The requirement to award 30 percent of subcontracts to start-ups and small businesses is not a mandatory requirement of the EU tender.
At the same time, the bundling of research tasks at a few corporations desired by Brussels ensures that SMEs or emerging start-ups cannot acquire any competencies at all in such important areas as crypto-communications, for example, or develop them further at all. In Germany, start-ups are concerned about whether the desired “impulse effect” of IRIS² for the European space industry will unfold or whether the “spillover effect” on the new space ecosystem in particular will completely fail to materialize in the current constellation.
Reflex Aerospace, for example, manages to design and produce a satellite in less than a year – while established manufacturers calculate up to four years for the development and production of satellites. With its disruptive approach, Reflex Aerospace is unlikely to be used as a pure supplier. Which member of a consortium will commission its direct competitor, which operates significantly faster and cheaper on the market with the aid of agile product development and additive manufacturing, if this group can guarantee to sell its own products more expensively within the framework of IRIS²?
In space, it becomes apparent that rapid development and inexpensive reproduction are important to keep a satellite constellation technologically up to date. Satellites today are renewed about every five years. Optimized manufacturing time also means being able to deploy improved technologies faster and communicate even more securely via satellites. Unfortunately, IRIS² is not exploiting its innovation potential here.
For Germany, in particular, much is at stake. An important industry of the future that is already creating high-tech jobs is being knowingly excluded from major public contracts. The European constellation does not benefit. It is tying itself to the traditional corporations for years to come – along with its business as usual and all the costs that the large consortium entails!
It is not too late to incorporate the New Space ecosystem, at IRIS²:
We need active policies in Berlin and Brussels to strengthen and expand Germany as a space location.
Dennis Moore is Vice President of Sales & Business Development at Reflex Aerospace.
Convinced Europeans and democrats can only be disturbed by what could be heard at the AfD’s European election meeting in Magdeburg over the weekend. Top candidate Maximilian Krah ranted about family albums and ethnicity, while co-party leader Alice Weidel wants to turn Europe into a “fortress.” Several successful candidates in the top positions had been supported by the party’s far-right camp.
Its figurehead Björn Höcke provoked with the statement: “This EU must die, so that the true Europe can live.” Allegedly, he wants a new European confederation. Also not off the table is the demand for Germany to leave the Union. The adoption of the election program could drag on into January; in any case, more candidates are to be nominated this Friday.
The divisive potential is also evident when looking at the AfD’s role models. The party presents itself as Putin’s mouthpiece, said CSU Secretary General Martin Huber. Krah is criticized for his closeness to China, and several candidates cited the right-wing populist Orban as a role model.
In the coming months, it will also be important to expose the language AfD politicians use when they talk about “globalists” or “immigration.” False statements are also likely to increase between now and the election date, as was already evident in Magdeburg: whether on the imagined “abolition of cash” or on misrepresented quotas from the asylum compromise. What is needed are thus alert eyes and ears and a clear European course.
All over the world, governments are grappling with how to harness the benefits of artificial intelligence while minimizing the risks of this powerful technology. Germany is relying on the EU, which is currently negotiating the AI Act. Once in force, this will also become directly applicable law in Germany. “We are not planning any national AI regulation in Germany,” says Anna Christmann (Green Party), the Federal Ministry of Economics’ Commissioner for the digital economy and startups, in an interview with Table.Media.
Volker Wissing’s (FDP) digital ministry also sees no reason for Germany to go its own way. “We want to back developers of AI in our country by simplifying access to data and creating clarity about future standards,” says a BMDV spokesperson. “Purely national rules are counterproductive here.”
The German government is thus involved in various international regulatory processes – with distributed responsibilities. The Digital Ministry is in charge of the G7 Hiroshima Process on generative AI, while the Economics and Justice Ministries are leading the negotiations on the AI Act in the EU. As an export nation, Germany is committed to international standards wherever possible “so that developments from Germany are compatible and vice versa,” says the BMDV spokesperson. There, it is important not to slow down developments “by imposing too many complicated rules. That is why Germany is campaigning at the G7 level “for clear transparency rules that leave room for innovation.”
Christmann also says: “We must not stifle development at this early stage. Otherwise, there is a risk that Europe will become unattractive for AI developers.” However, he says it is also clear that we need guardrails for AI that affect fundamental rights. “Something like social scoring or warrantless biometric identification in public spaces has no place in a democracy.”
Christmann has just returned from a trip to the US, where she met AI researchers and entrepreneurs, among others, in California. “In the US, there is no understanding of regulating this innovative technology so harshly,” she says. “We need to balance regulation with encouraging research and development.” She says it is also crucial to keep in mind that AI companies in Europe are just emerging, so Europe has different conditions than, say, the United States. Nevertheless, “We must not always look enviously at the USA. Our startup scene is improving.”
Christmann sees it ambitious that the Americans have not waited until the G7 or the EU have agreed on guard rails or rules for AI but have instead brought their companies to a voluntary commitment. “It’s important to be in exchange with the big technology companies, precisely to avoid bias.” And it is right to put the already existing models on a reasonable footing, she said. “But I don’t share the opinion that a voluntary commitment alone is sufficient,” she stresses.
After the summer break in Brussels, the trilogue negotiations enter the hot phase. The aim is to conclude them under the Spanish Council presidency. Christmann also believes it is important to conclude the negotiations on the AI Act promptly to have reliable regulation. “As long as it is unclear what is coming, investors are holding back. That’s not a good situation.” The law must be ready before the European elections, she said.
Germany had made some comments on the general direction of the Council, including that only relevant and proportionate requirements should apply to general purpose AI (GPAI) systems. And pointed out that the review was not yet completed. In fact, this was before the discussion on ChatGPT reached the public.
That means the discussions – including in the Council – are not over. “We have to ask ourselves whether regulation is fulfilling its purpose of shaping technology for the benefit of people or whether it is stifling development,” Christmann cautions. It is good to limit regulation to high-risk applications, she says. At the moment, however, she says, Europe is all about increasing detailed regulation. “That leads to more difficult conditions, especially for young AI companies like we have in Europe.”
For example: “I cannot see why companies should be required to carry out an additional assessment of the impact of fundamental rights, as proposed by the EU Parliament,” criticizes Christmann. Because that would be time-consuming to impossible, especially for young companies. However, it was the Greens/EFA group that introduced the Fundamental Rights Assessment into the legislative proposal.
What else is important for Germany in the coming negotiations? “To create good development conditions for SMEs and startups,” says Christmann. “We need experimental fields. It must be possible to try things before full regulation takes hold.” That is why, she says, the German side attaches importance to the establishment of real labs and that there are time exceptions for SMEs.
Another point is the creation of exceptions within the high-risk areas. So that, for example, cleaning robots in hospitals are not subject to the same security requirements as patient records, for example. “Here we are working on solutions, the discussion is still ongoing,” Christmann said.
Since the Council had already decided on its General Approach before the discussion on Foundation Models or generative AI (such as ChatGPT) gained momentum, member states had hardly shed any light on the topic. “It is right that the EU Parliament has taken up Foundation Models,” Christmann says. But the German government is still in talks about which regulation is appropriate. The framework conditions have changed much more than usual since the Council position was adopted. Therefore, the need to work on concrete solutions is greater.
Digital Minister Wissing has already made a proposal in the FAZ: He envisions mandatory self-regulation for generative AI. However, he also warns against leaving the formulation to the large US tech companies alone. In addition, there should be a review after just two years. Then it would be possible to assess whether a code of conduct for generative AI is sufficient as a supplement to the AI Act – or whether tough regulation and integration into the law are needed.
Meanwhile, the discussion about the implementation of the AI Act has not yet begun. But Christmann can already say one thing about it: “We want to avoid the mistakes of the GDPR and not have 16 different interpretations of the rules again.” The solution could be a central office, an AI Office – something like the one proposed by the EU Parliament. But there is a need for clarification there, too. “A European AI Office must not become the eye of the needle for AI approval throughout Europe,” Christmann warns.
Professor Uszkoreit, you recently founded a German ChatGPT start-up with a presence in China. What is the objective?
We established the start-up Nyonic in Berlin, with a research team in Shanghai. Our aim is to address two weaknesses of ChatGPT. Current language models are predominantly focused on English and perhaps Chinese. We intend to change that. Additionally, we want to transform ChatGPT into a reliable work tool for specific industries, which it currently isn’t.
But why different languages? English is a global language, and Chinese is heading in that direction.
Much European knowledge is not yet available in English, but only in various European languages. Additionally, AI systems should bring the same benefits to users worldwide. Language diversity is a significant step for ChatGPT and crucial for a multipolar world. The majority of the world is multilingual. The US can afford Anglo-centric technology, and China can afford China-focused technology. However, the majority of the world, including Southeast Asia, Africa, and Europe, relies on covering numerous languages and cultures.
Why do you take on this effort? At 73 years old, you are already at retirement age. You are and remain a pioneer in European AI. You could relax.
Everyone on our team, regardless of age, wants to be part of the biggest AI revolution of our lifetime. Additionally, we wish for Europe not to be left behind by the US and China this time. Europe possesses many intelligent individuals, and numerous innovations have come from here. But then we repeatedly failed. China already has more than 80 language models similar to ChatGPT, while the EU only has three. In Germany, there’s only one, which is not yet among the best. We aim to change that.
The topic of artificial intelligence is not new. You have been working on it for over 30 years. What’s different about ChatGPT that has sparked so much excitement now?
The system has achieved something we call understanding, similar to humans. It’s defined differently for the machine since it lacks our holistic worldview and doesn’t adapt its knowledge based on user input. But the effect is the same: The machine responds as if it truly comprehends the inputs.
What are ChatGPT’s remaining weaknesses?
Although ChatGPT is exceptionally intelligent, it still exhibits surprising knowledge gaps that it occasionally fills with imagination. We cannot fully rely on its knowledge and judgment yet.
Will the next version fix these issues?
You might be surprised to learn that there’s currently no reliable method to train the machine to get exactly what I want. It’s all still trial and error, more of an art than a science. There’s no curriculum; I simply feed data into giant tubes and see what happens. We don’t truly know what the machine is doing in the process. When the results come, we say: Well learned, but dear machine, you shouldn’t believe everything you’ve read. And don’t say everything you believe. Then we specifically feed data to correct or improve the machine’s behavior and see what happens.
AI can be misused by malicious people. That needs to be prevented, right?
Yes, we need rules, especially for the specific critical uses of AI. However, those who design these rules should resist the temptation to regulate such a complex technology as a whole. By the time the law takes effect, AI will have progressed further. Currently, I am more concerned about excessive regulation than evil AI systems. This danger is much greater in Europe than in the US or China.
What is the consequence?
Even research aimed at taming and controlling increasingly powerful AI through appropriate control technologies could be hindered by excessive regulation. The most dangerous are always the meta-regulators who want rules to apply universally.
But Germany has a constitution, a basic law.
It doesn’t make sense to invent a kind of basic law for a new technology that is rapidly evolving when we don’t yet know where it’s heading. It is much more sensible to regulate AI for specific application areas. For example, AI for driving, AI for the pharmaceutical industry, for aviation, or schools. This not only makes sense but is also crucial. Each area has very different application conditions. The Chinese do it this way too. Then, in the distant future, we can distill a kind of AI constitution from the intersection of these individual regulations.
AI will change the world as profoundly as the invention of electricity. If I say we need central regulations for the use of electricity, you immediately realize that’s nonsense. Or the steam engine. If I use it in a factory, it requires different rules than if I use it in a locomotive. AI remains a tool of humans and is not a competitor to humans.
Why is this fear, especially in Europe, particularly in Germany and France, so pronounced compared to Asia and the USA?
That is a good question, and it is essential to acknowledge this fact. Perhaps it is a desperate attempt by Europeans to compensate for their waning influence in the world by trying to dominate the meta-level with their rules. The stronger desire for regulation has grown as they no longer have world powers. In Europe, especially in Germany, some medications require prescriptions, while in the US, they can be bought at the supermarket. In Europe, the focus is more on improbable exceptions than on daily realities. There seems to be almost a great pleasure in worrying about risks rather than bringing technology to success through skill and full commitment.
In the US and China, they pragmatically focus on the most likely scenarios.
In the EU, the division of political tasks between member states and centralized regulation has led to a very particular development. Instead of aiming for a modern, flexible combination of codified law and case law that suits the dynamics and diversity of our continent, this distribution of power has pushed the European continental legal system of codification to the extreme with a mammoth effort.
This tendency combines with the hope that legislation can rid the world of evil and impose one’s own ethics through regulation. This is now holding us back, and the world, led by the BRICS countries, is no longer interested in adopting the rules. Furthermore, fear is being fueled more in Germany than in other countries. In the meantime, a new profession of self-proclaimed AI ethicists, AI lawyers and algorithm hunters has emerged. They make a living by spreading fear. These are people who, strictly speaking, do not know much about AI. I am also unsure if they are the best jurists and social scientists in this field.
The other extreme would not be favorable either: every application needs separate certification because little is centrally regulated.
That is the concern of SMEs and particularly of craftspeople and small businesses. If they have to certify everything separately, it takes so long and becomes so expensive that SMEs say, “It’s not worth it for us.” So, we must limit the certification requirements to the truly critical areas and burden producers and users with the burden of proof only where there are actual exceptions and reasons for concern. And provide cost-effective testing procedures with the help of technology. Above all, we must consider what is practical in everyday life. I am now very curious whether Germany and the EU will get it right this time or act as if they can still determine the rules of the world.
Hans Uszkoreit, 73, has worked in Germany, the US and China as a university professor, research manager, industry advisor, and co-founder of several start-ups. He is considered one of Europe’s leading AI researchers. Uszkoreit is the scientific director at the German Research Center for Artificial Intelligence (DFKI). He initiated and coordinated German and international research collaborations and led several renowned European projects. Uszkoreit has authored over 250 international publications. Recently, he co-founded the Berlin-based ChatGPT-like start-up “nyonic” together with a research team in Shanghai. His wife, Xu Feiyu, is also a prominent AI researcher and was previously the head of AI at SAP.
*The second part of the interview can be found in China.Table.
The EU Commission wants to establish a European satellite constellation. The IRIS2 program – Infrastructure for Resilience, Interconnection and Security by Satellites – stands for securing Europe’s sovereignty. Between 2025 and 2027, up to 170 LEO satellites are to be placed in low Earth orbit. More than €3 billion in public funding is being provided by Brussels and the European Space Agency (ESA).
Only European companies are to be involved in the program. The aim is to be able to act autonomously in the event of a crisis. Thierry Breton, the Internal Market Commissioner responsible for 2022, emphasized: “This is historic! After Galileo and Copernicus, we are adding a third constellation to our European portfolio of strategic space infrastructures! […] IRIS² will be a New Space constellation … integrating the know-how of the major European space industries-but also the dynamism of our start-ups, who will build 30% of the infrastructure!”
As a start-up building innovative dual-use satellites, Reflex Aerospace had high hopes for IRIS². But already during the EU legislative process, concerns arose that the program would primarily serve France’s interests and its well-positioned space corporations. Germany, on the other hand – as the largest contributor to both the EU and ESA – advocated from the outset for the participation of SMEs (small and medium-sized enterprises) and companies from the fast-growing new space ecosystem.
This comprises young companies with fresh ideas that, thanks to their optimized processes and improved products, challenge the traditional players in space. In Europe, these are mostly large (semi-) state-owned companies. In Germany, in particular, a structurally significant new space industry has emerged. By means of micro-launchers, its own satellite construction and laser communication, it is in a position to build its own constellation!
It was explicitly intended to involve SMEs and new space companies in IRIS². At least, that is what the Regulation of the European Parliament and of the Council of the European Union of March 15, 2023 (EU 2023/588) says: “The program [IRIS² should] maximize the exploitation of innovative and disruptive technologies as well as novel business models developed by the European space ecosystem, including New Space, in particular by SMEs, mid-cap companies and start-ups …”. For subcontracts worth more than €10 million, the goal is to “ensure that at least 30 percent of the contract value is subcontracted through competitive bidding at various levels to companies that are not part of the prime bidder’s group…”!
So much for the theory. Unfortunately, “the little ones” were unable to apply directly in the bidding process for the main part of the constellation. The process started on March 23, 2023, and one of the conditions for participation set by the EU Commission was a minimum annual turnover of €500 million. As early as May 2, 2023, Europe’s established space groups – including Airbus Defence and Space, Eutelsat, Hispasat, OHB, SES, Telespazio and Thales Alenia Space – announced the formation of an “open” consortium. This concentration on the long-established space companies – without competition from a second consortium and excluding New Space – is dangerous for Europe’s innovative strength!
Under IRIS², corporations are now able to bypass and keep down unwelcome competitors with appropriate subcontracts. Even subcontracts from new space companies or SMEs remain optional. The requirement to award 30 percent of subcontracts to start-ups and small businesses is not a mandatory requirement of the EU tender.
At the same time, the bundling of research tasks at a few corporations desired by Brussels ensures that SMEs or emerging start-ups cannot acquire any competencies at all in such important areas as crypto-communications, for example, or develop them further at all. In Germany, start-ups are concerned about whether the desired “impulse effect” of IRIS² for the European space industry will unfold or whether the “spillover effect” on the new space ecosystem in particular will completely fail to materialize in the current constellation.
Reflex Aerospace, for example, manages to design and produce a satellite in less than a year – while established manufacturers calculate up to four years for the development and production of satellites. With its disruptive approach, Reflex Aerospace is unlikely to be used as a pure supplier. Which member of a consortium will commission its direct competitor, which operates significantly faster and cheaper on the market with the aid of agile product development and additive manufacturing, if this group can guarantee to sell its own products more expensively within the framework of IRIS²?
In space, it becomes apparent that rapid development and inexpensive reproduction are important to keep a satellite constellation technologically up to date. Satellites today are renewed about every five years. Optimized manufacturing time also means being able to deploy improved technologies faster and communicate even more securely via satellites. Unfortunately, IRIS² is not exploiting its innovation potential here.
For Germany, in particular, much is at stake. An important industry of the future that is already creating high-tech jobs is being knowingly excluded from major public contracts. The European constellation does not benefit. It is tying itself to the traditional corporations for years to come – along with its business as usual and all the costs that the large consortium entails!
It is not too late to incorporate the New Space ecosystem, at IRIS²:
We need active policies in Berlin and Brussels to strengthen and expand Germany as a space location.
Dennis Moore is Vice President of Sales & Business Development at Reflex Aerospace.
