Who Controls AI?

In recent months, doomsday scenarios and calls for tighter regulation have increasingly shaped the debate around artificial intelligence. When AI agents developed by OpenAI went beyond their intended task in July and attacked another platform's system, concerns that AI could ultimately pose an existential threat to humanity suddenly took center stage. In a rare show of agreement, influential AI leaders including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, as well as Elon Musk, have called for slowing down the development of the most powerful AI systems.

What is behind these calls? How can Europe chart its own course between technological progress and social responsibility? And what role can foundations play? Kerstin Lohse-Friedrich, Vice President Communications at Robert Bosch Stiftung, spoke with tech journalist and podcast host Fritz Espenlaub, who will deliver the keynote on "Human. Power. AI." at the Robert Bosch Forum in Stuttgart.

Mr. Espenlaub, for a long time, the major AI companies were primarily focused on developing as quickly as possible and pursuing rapid growth. Why are some of the industry's leading figures now calling for greater regulation, and in some cases even for slowing down development?

Fritz Espenlaub: Fear of superintelligent AI has been around for as long as the current AI revolution itself. It is the flip side of the enormous hopes that many in the AI industry have placed in this technology. For key players in Silicon Valley, the thinking is clear: we are working on humanity's last invention, one that will either solve all our problems or ultimately grow beyond our control. It is this latter concern that is driving the current debate around regulation.

How sincere the tech leaders are in their calls for regulation, however, is open to debate. The leading AI labs could also benefit from regulation if it were to make it harder for new competitors to enter the market. That makes the debate particularly complex.

When Sam Altman, Dario Amodei or Elon Musk talk about regulation, what exactly do they mean? Government oversight, technical safety standards, or something else?

All three see themselves caught in a prisoner's dilemma: each believes they cannot afford to slow down their own development because their competitors would overtake them. Government regulation, or international agreements, could theoretically resolve that dilemma. What remains unclear is what this would look like in practice.

One recent proposal from Anthropic CEO Dario Amodei is that the leading AI companies should submit to permanent oversight by external experts and give them access to all internal developments.

AI has long since become a geopolitical issue as well. While the United States and China are investing enormous sums and competing for technological leadership, Europe is struggling to find its place. Is there any realistic prospect of internationally agreed rules for AI?

This is certainly not the best moment for international agreements. But I would not rule them out. Even during the Cold War, the United States and the Soviet Union managed to find common ground on certain issues. In principle, that should also be possible in today's tense geopolitical environment.

European researchers and policymakers recently presented an AI strategy for Europe aimed at ensuring that the continent does not fall behind. How do you assess the strategy?

There is a lot to commend in it. One important insight, for example, is that Europe must secure access to frontier AI at all costs and understand this as a geopolitical interest. But our big problem is that we have a great many strategy papers and still far too few data centers and AI unicorns of our own.

IYour forthcoming book is titled "Das Ende der Welt ist erst der Anfang" ("The End of the World Is Only the Beginning"). In it, you ask how technological developments can contribute to freer and more dignified lives. What would it take to achieve that?

First, we need to recognize the full significance of this historic moment. The stakes could hardly be higher for Europe, and we will not be able to preserve our freedom, independence and prosperity unless we finally start taking this challenge truly seriously. At the same time, we should never lose our enthusiasm for progress. The transformations we are witnessing also offer an opportunity to carry European values into the future and become more self-determined.

Foundations are also working toward a positive vision of the future. The Gates Foundation plans to spend an additional US$1 billion over the next two years to democratize AI in countries of the Global South. Where do you see the greatest potential for foundations to make a difference?

Throughout human history, we have solved a great many problems through good ideas. In principle, today's problems can also be solved if we come up with the right ideas. And where do ideas come from? From human minds. So, in my view, the greatest lever philanthropic organizations have is to enable more people to develop good ideas, whether by funding research or simply by giving more people access to education. The rise of AI does not change that. Quite the opposite!

What would a positive AI scenario for Europe in 2035 look like to you?

AI is safe and powerful, and it helps us create the foundations for a good life. European AI is one strong ecosystem among others, resilient and operating at the technological frontier. The possibilities that could emerge from this are almost limitless. But there is still a great deal to be done before we can make that scenario a reality.