EU Parliament Approves Changes to AI Act, Sparking Debate Over Regulatory Approach
Berlin, March 18, 2026 – The European Parliament’s Committees on the Internal Market and Consumer Protection (IMCO) and Civil Liberties, Justice and Home Affairs (LIBE) today adopted their position on adapting the EU AI Act. The vote, passed by a majority, supports a shift in how certain AI regulations are applied, potentially removing some high-risk product areas from the Act’s direct scope. Instead, these areas – including medical devices, machinery, and toys – would be regulated through existing sectoral legislation.
A Shift Away From Horizontal Regulation
This proposed change represents a significant departure from the AI Act’s original intent as a comprehensive, horizontal legal framework for artificial intelligence. According to the TÜV Association, this move could weaken the EU’s level of protection for high-risk AI systems. Dr. Joachim Bühler, Managing Director of the TÜV Association, stated, “If the European Parliament prevails in the negotiations with the member states and the EU Commission, the European path to artificial intelligence will end in a dead end.”
Concerns Over Fragmentation and Implementation Delays
Critics argue that shifting regulation to a sectoral approach would fragment the European legal landscape for high-risk AI. Requirements for AI would then need to be redeveloped and integrated into numerous existing regulations, potentially leading to differing standards, procedures, and implementation deadlines. This could similarly significantly slow down the regulation of high-risk AI, as AI-specific requirements would need to be laboriously incorporated into various sectoral laws and continually updated.
Given the rapid pace of technological development, adapting AI requirements across numerous sectors in a timely manner could prove challenging. Outdated regulations could then become an obstacle to innovation. A horizontal approach, proponents argue, can better keep pace with AI’s dynamics while maintaining a high level of protection.
“That wouldn’t be a reduction in bureaucracy, but rather the entry into a regulatory patchwork with more effort, more uncertainty and higher long-term burdens for companies,” Bühler added.
Impact on Global Standards and Competitiveness
The AI Act has been viewed as a potential global standard for trustworthy AI, with many international actors looking to the European approach. By softening its regulatory stance, the EU risks losing influence on international standards, which could have economic consequences. International standards dictate the requirements that prevail in global markets.
“If Europe waters down its approach to high-risk AI, we will weaken our influence on rules that will apply worldwide tomorrow,” Bühler warned. “And that also weakens Europe’s competitiveness in the global AI competition.”
Next Steps: Trilogue Negotiations
The TÜV Association, along with a coalition of organizations including AlgorithmWatch, AI Policy Lab, ALLAI, DEKRA, ForHumanity, Open Knowledge Foundation, The Future Society and TÜV AI.Lab, has already voiced concerns about a sector exit in an open letter to the federal government. The upcoming trilogue negotiations between the EU Parliament and member states will be crucial in defending the horizontal approach of the AI Act and ensuring the continued security and trust in high-risk AI systems.
“Reversing the AI Act through the back door would be the wrong approach. Instead of removing central product areas from the legal framework, interpretative guidelines from the European Commission are needed quickly to clarify overlaps with sectoral law and to ensure a high level of security,” Bühler stated.
Upcoming MeetUp: EU AI Omnibus – What’s Next with the AI Act?
On March 19, 2026, the TÜV Association will host a MeetUp titled “EU AI Omnibus: What’s next with the AI Act?” The event will feature discussions with Axel Voss MEP, Parsa Marvi MdB, Anisa Idris, board member of the National Association of Digital Health Care eV, Prof. Dr. Annette Zimmermann, University of Wisconsin, and Franziska Weindauer, CEO TÜV AI.Lab, regarding the European Commission’s plans to adapt the EU AI Act as part of the AI omnibus. The online event will take place from 9:00 a.m. To 10:00 a.m.