European AI Act & AI Ethics – What AI Startups Need to Know
The EU AI Act is the world’s first comprehensive set of regulations for artificial intelligence and is intended to minimize risks, promote innovation and protect fundamental rights. The regulation is gradually coming into force, so there are still many question marks over its concrete implementation. The “European AI Act & AI Ethics” event organized by AI Startup Rising | hessian.AI offered an in-depth discussion of the current status of the regulations and the potential impact of the EU AI Act on startups and the AI ecosystem in Europe.
After a short introduction to the event by AI Startup Rising co-lead Carlina Bennison, four expert inputs followed, which dealt with ethical questions and regulatory challenges. Afterwards, a high-profile panel with Manuel Brack (DFKI | TU Darmstadt), Prof. Dr. iur. Janine Wendt (TU Darmstadt) and Anna Spitznagel (trail), moderated by coach Stefan Rötzel, discussed the opportunities and risks of the AI Act, particularly with regard to European competitiveness. AI start-ups from the Accelerator RISING and AI Startup Competition formats also contributed to the discussion with questions and experiences.
Impulses from the Experts
Manuel Brack (AI Researcher, DFKI | AIML Lab | TU Darmstadt | hessian.AI | The Hessian Center for Artificial Intelligence)
Manuel Brack presented his co-authored paper “The Cake that is Intelligence. And who gets to bake it”. Using a vivid cake analogy, he looked at ethical challenges along the value chain of large language models: from the “ingredients” (data) to the “recipe” and the “baking process” to “tasting and selling”. He made it clear that errors and risks can occur in every phase, from data collection to training and operation, and that comprehensive quality management is necessary. His appeal to developers: Ethical issues should be integrated into the development process right from the start.
Prof. Dr. iur. Janine Wendt (hessian.AI | The Hessian Center for Artificial Intelligence , TU Darmstadt)
As an expert on the EU AI Act, Prof. Dr. Janine Wendt provided a detailed classification of the new regulation. She emphasized the risk-based approach, which does not treat all AI products equally, but differentiates between high-risk and low-risk AI as well as between AI operators and AI users. A key problem is that many details of the interpretation are still unclear and will only be fleshed out in the coming months. Wendt therefore called on start-ups to actively voice their needs via associations and networks in order to ensure a balance between regulation and the promotion of innovation.
Patrick Aurin (Federal Network Agency, via video message)
Patrick Aurin, an expert on artificial intelligence at the Federal Network Agency, spoke via video message in which he emphasized that the AI Act is an opportunity for Europe to position itself as a pioneer for trustworthy AI. At the same time, he pleaded for bureaucracy to be kept to a minimum and for innovation not to be slowed down. He also called on start-ups to play an active role in shaping the implementation guidelines.
Anna Spitznagel (CEO, trail)
Anna Spitznagel contributed the startup perspective. Her young company trail deals with governance and regulations – not only in connection with the AI Act. She could understand the concerns of start-ups that increasing bureaucracy could impair agility, but also pointed out that product safety requirements already exist in many industries and often contribute to quality assurance. At the same time, she saw a great opportunity for European companies to strengthen their trustworthiness on the international market by complying with the AI Act.
Panel discussion: Opportunities and risks of the AI Act
In the panel discussion moderated by AISR coach Stefan Rötzel, the question of regulatory balance was at the center of the discussion. The experts agreed that regulations are needed for new technologies such as AI. At the same time, there is a risk that excessive bureaucratic hurdles or ambiguities could impair innovation and competitiveness.
Manuel Brack emphasized the excellent quality of AI research in Europe. However, it is still too rare to translate research results into marketable applications. hessian.AI | The Hessian Center for Artificial Intelligence is a positive example here; there are still few research and innovation centers that provide researchers with the necessary infrastructure and associated services for AI research projects quickly and easily.
With regard to start-ups, it was emphasized that European start-ups would face challenges even without the EU AI Act; low venture capital investments, a fragmented market, the lack of global European big tech players and strict data protection rules were cited as hurdles in the discussion.
Despite some critical voices, a positive attitude towards the EU AI Act prevailed in the discussion. Many of the participants saw it as an opportunity to promote ethically responsible AI development and position Europe as a pioneer for trustworthy AI. Particularly in light of current political developments, it is important for society to combine technological progress with democratic values.
Conclusion: both a challenge and an opportunity
The EU AI Act presents European AI startups with new regulatory challenges, but as a “seal of quality”, the regulation also offers the opportunity to position themselves as pioneers for safe, ethical and trustworthy AI. The event showed that the European AI scene is vibrant and engaged and that startups can play an active role in shaping the AI Act.
The startups and experts in attendance agreed that Europe has the potential to become a global leader in the development of responsible AI – provided that the framework conditions strike the right balance between regulating risk and promoting innovation.
