Summary
Today’s AI still relies on people to work alongside it, control and guide it. What specific political responsibilities come with ensuring good governance in light of this new technological advancement? We spoke with an expert in public administration modernization and innovation about how AI is transforming the state.
Lars Zimmermann is a co-founder and board member of GovTech Campus Deutschland e.V., the world’s first innovation, development, and learning hub dedicated to modernizing government and administration. The German government is promoting the Campus as a collaborative platform aimed at developing digital innovations and technologies for federal, state, and local administrations in partnership with the tech community. The goal is to make these advancements available for reuse.
Before founding the GovTech Campus, Zimmermann worked as a technology and transformation consultant and was the founder and board spokesperson of the Stiftung Neue Verantwortung. He has been engaged for many years in state modernization and administrative reform, developing numerous initiatives and projects in this field. Since early 2024, he has been working as a research associate at the Konrad Adenauer Foundation.
—With the rise of AI, are we on the verge of a societal leap comparable to the arrival of the steam engine?
—Honestly, we don’t know yet. I’m always a bit cautious about declaring a major revolution. However, I suspect that AI will have a similarly profound impact on society. We are on the verge of a development leap that would have been unthinkable 15 years ago. Just four years ago, none of us were even discussing, let alone working with, large language models like ChatGPT.
There has been a significant disruptive breakthrough in AI that has, for the first time, reached the general population at various levels. If we extrapolate from this advancement, we must conclude that AI, much like the steam engine or other groundbreaking technologies, could represent a significant leap forward in development. Of course, this comes with risks, but above all, there are also numerous opportunities. In my opinion, this evolution has arrived at the right moment.
—Why did it arrive at the right moment?
—For example, globally, especially in industrialized countries, we face the challenge of a growing demographic gap. If we develop and utilize artificial intelligence effectively, we can help address these shortfalls in many professions. A second example is the enhancement of computing capabilities. With artificial intelligence and the strengthening of IT infrastructures, we can now process, manage, and contextualize data—capabilities that were not available to us ten years ago. These advances bring significant benefits, particularly in health research and safety matters.
So, I believe that while we are in a time of significant challenges, we also have tremendous opportunities because of AI. We are witnessing major technological advancements that can help us combat diseases, develop medicines, and address workforce shortages.
—What time frame are we discussing now?
—It’s a good question that no one can answer with certainty today. Perhaps we can frame it this way: short-term innovative leaps in AI are often significantly overestimated, while those in the medium and long term are greatly underestimated. If we look back, four years ago, nobody knew what a large language model was. I am confident that we will witness significant leaps in development over the next 15 to 20 years.
—And in the short term?
—In the short term, for example, large language models can lead to efficiency gains. These models are already quite powerful and capable of handling many tasks. However, they cannot, for instance, replace advisors in a federal ministry. Today’s AI still needs people to work alongside it and to oversee and guide its operations. Clearly, this could change when technology and computing capabilities advance to the point that AI can take on certain tasks without human assistance.
Naturally, these changes will also impact us. AI could assume tasks currently performed by humans, making the corresponding jobs superfluous. Despite this, I do not believe that AI will result in mass unemployment. On the contrary, I see AI as a means to address gaps created by labor shortages. It is not about reducing resources but rather about filling those gaps.
—AI is often viewed as a threat and is frequently linked to job losses, fake news, and deepfakes. So, what are the opportunities?
—For one thing, AI can significantly enhance administrative processes. It offers administrative staff a tool that allows them to perform their tasks more efficiently and effectively. A good example of this is large language models, which enable texts to be written and synthesized quickly.
Suppose that ministerial advisors need to create templates. Often, a minister will ask for a long text to be condensed into a concise page. In the past, this used to take a considerable amount of time. With today’s large language models, this process can now be easily automated. What once took an hour can now be completed in just a few seconds. This saves a significant amount of time and boosts efficiency.
Another example is this very interview, which we conducted in German but will later be printed in Spanish. In the past, translation would have taken a lot of time and money. Today, thanks to large language models like ChatGPT, texts can be translated quickly and efficiently into any language. This considerably simplifies and accelerates the entire process.
—However, we cannot replace the translator we currently employ. We still need someone to review the solutions generated by the language models.
—I don’t believe it’s wrong for a human to continue overseeing translations. However, it is likely that translation technology will advance to the point where human oversight may no longer be necessary. The error rate could then be comparable to that of a simultaneous translator, who is also not error-free. This development is likely to occur relatively quickly. However, I do not believe that translators will be out of work. When it comes to analyzing and interpreting speech, as well as assessing nuances that indicate self-confidence or insecurity in the voice, AI is not yet ready. These skills require human perception, so it will take time before AI reaches the same level in this area. However, it is very likely that it will get there eventually.
—Does this not imply that primarily routine tasks requiring average qualifications will be replaced, thus eliminating many jobs, while a shortage of highly qualified specialists remains?
—I’m not sure about that yet. I believe mid-level positions are more at risk than very low-level ones. The more demanding a job is—meaning mid-level or higher skilled—the more likely it is to be replaced by AI. This also impacts many oversight roles in administration that are currently handled by more qualified specialists but could be automated in the future. For example, positions held by academics in the insurance industry could also be at risk from AI. So, it’s not just basic jobs, but also those that require a higher level of qualification. In the next phase, AI may not make any job obsolete, but it will alter the traditional requirement profiles for those positions. Those who do not adapt to this change will find it challenging.
—We would like to hear your thoughts on how the development of AI will impact Latin America.
—Ultimately, the development of AI is still evolving. It is less about how proficient different countries are at developing AI and more about the capabilities of companies in this field. Currently, only a handful of companies are driving AI on a global scale, and most of them are based in the United States.
The impact of AI on various social systems remains unclear. Opportunities and risks are present in industrialized countries as well as in other parts of the world. Take Africa, for example: many African countries have made greater progress than Germany in payment systems and digital support for microenterprises. This demonstrates that even less developed economies can benefit from technological advancements.
If economies in South or Central America continue to develop, there is a real risk that AI-supported innovations will also lead to increased labor efficiency and, consequently, job cuts. This means that governments in these economies must carefully examine the impact of AI to avoid being left behind and to make the most of the opportunities presented by this technology.
Governments
—Currently the largest companies in the AI sector come from the United States. The so-called network effect is particularly strong in this field, where the market leader gains significant advantages and leaves little room for competitors.
—There’s a well-known phrase: “In the field of artificial intelligence, there is no third place anymore.” You’re either the world market leader or the one trying to catch up. But this risk applies to all countries. In my view, this is not a typical issue of development politics.
It’s important to understand that these challenges cannot be tackled by a single country alone. Especially in regions like South America, it would be crucial for countries to collaborate at the supranational and regional levels to effectively address the issue of AI. However, good governance remains essential. Poorly governed countries will continue to struggle to leverage technological innovations effectively. High-quality governance rooted in fundamental democratic values will be even more crucial in the face of rapid technological progress. AI cannot take on the task of good governance; this responsibility must stay in the hands of citizens and will likely become more important.
—What specific tasks should politics address to ensure good governance in light of these new technological advances?
—First and foremost, I believe politicians need to cultivate a strong curiosity about AI. Countries that thrive in times of innovation are often those that embrace new technologies with openness and curiosity. It’s important not to fear or demonize technology. Openness to new developments is crucial.
Second, cooperation is essential. No country, not even the United States, can tackle AI alone and with the necessary depth. Even large nations like the United States and China face significant challenges. For smaller countries like Germany and France, cross-border collaboration is even more crucial. Therefore, a strong commitment to cooperation and working together on these issues is even more important.
The third step is also very important: creating the framework conditions that enable innovation.
The fourth step is the actual application of AI. It’s important that countries and their citizens do not demonize AI but actively engage with it. Those who use AI also have an opinion. An example from the past: the Germans became world leaders in the automotive industry because the domestic market was strong, and Germans themselves bought cars in large numbers. In the field of AI, no country has yet established itself as a leader in its application, whether in education or healthcare. Therefore, all regions still have similar starting points. In fact, countries that are not as saturated and have fewer existing structures may even have an advantage, as they could be more open to new developments.
The fifth step focuses on training. Education is crucial for all areas, including AI. The development and application of AI and other technologies will not succeed without top-tier training centers and educational systems. It is essential to have strong educational institutions that can prepare the next generation for new technologies and integrate them into the education system.
—What positive narratives could political parties promote about AI?
—There are many positive stories that can be told about AI. For example, it can help address our demographic challenges by making public services more efficient. It could help reduce the need for staff, which in the long run would lower administrative costs. It could also lead to better management of taxpayers’ money and potentially even lowering taxes.
Another advantage is the boost in innovation. AI enables the processing of large amounts of data and makes efficient use of data centers. This can lead to groundbreaking (disruptive, revolutionary) research in the healthcare sector. I am convinced that I will live to see cancer no longer be a death sentence. With advancements in computing power and health data processing, we could develop treatments that defeat diseases once considered incurable twenty or thirty years ago.
Consider the example of the war in Ukraine. Without AI capabilities, Ukraine would not be able to use its drones as efficiently and effectively for defense, despite limited access to traditional weaponry. This AI capability, combined with drones, is crucial to the country’s defense efforts.
These examples show that AI can bring breakthroughs across a range of political domains. Essentially, we can identify two broad areas: AI enhances efficiency and reduces costs, and it drives innovation and solutions in various fields.
—Will AI also lead to new forms of participation in a representative democracy?
—Before I answer, I’d like to mention that you’re speaking with someone who strongly supports the traditional form of democracy. I am convinced that technology will not relieve us of democracy’s essential tasks. What do I mean by this? I don’t believe that AI will automatically make us better democrats or more informed voters. In my view, democracy is a system that functions through people and institutions. People will always be the central actors in a democracy, making decisions through majority vote. That’s why I consider myself conservative, in the best sense of the word. I don’t believe that AI-driven democratic systems will save democracy.
—In the U.S. election campaign, we saw that voters are becoming increasingly transparent to strategists through the analysis of large data sets. What impact does this have on democracy?
—That is not necessarily a problem. Strong democrats make strong democracies. You allude to the fact that parties now understand their voters better and can place targeted ads—but this is part of an ongoing development. If we look back at economic history, the invention of the printing press allowed handbills to be printed and distributed in mass for the first time, and I’m certain the first political handbill appeared not long after.
The principle remains the same: technologies enable us to spread information faster and in real time. Is it risky for politicians to know more about the population? Not necessarily. It becomes dangerous when states have access to everything, with the power to consolidate this scattered information into a comprehensive data block and use it to shape policy. Here again, good governance will ultimately be the decisive factor.
AI could be indirectly blamed for this, because without it the state would not be able to collect this information to the same extent. But you have to be able to resist innovation. Democracy has survived the printing press and the invention of television, and it will also survive AI.
—What can people do to prepare?
—For one, it’s important to stay curious and open-minded. As long you engage with new developments, you’ll stay better informed and more self-confident. The biggest danger is becoming complacent and letting others constantly tell you what to think.
It’s crucial to question things. For example, why do you see ads for a car on Instagram after talking about it with your friends – Is the smartphone listening? Can it recognize those sentence fragments? Asking these kinds of questions helps you stay critical of new technologies and make informed decisions.
It’s important to develop a critical and constructive attitude. There’s no value in immediately rejecting everything new or accepting it without question. Instead, try to form a balanced opinion and critically analyze what’s new.
It’s also useful to try things out. When you understand how something works, you can make a more informed decision about whether to use it or not. Germany has gained innovation leadership in many areas because people were willing to experiment with new ideas. The same approach should apply to AI—people should explore and understand different applications.
For example, we’re already using AI today, perhaps without even realizing it. My bank uses a simple AI system to detect suspicious transactions in my account. This type of AI protects us and makes many everyday tasks easier.
To sum up, openness to new ideas, critical thinking, and practical experience are essential to find a path in the AI era and recognizing its opportunities and risks it presents to each individual. In doing so, everyone can help move society and the country forward on this path.
German-Spanish translation by Manfred Steffen
