What’s Covered?
The paper focuses on harmonized technical standards and their legal function within the EU AI Act. It examines how CEN/CENELEC and ISO/IEC work together on developing standards and what practical hurdles lie ahead, especially for implementation. The authors cover:
• Legal and regulatory context of AI standardization in the EU
• Role of harmonized standards and the presumption of conformity
• Interplay between EU institutions, national bodies, and global SDOs (Standards Development Organizations)
• Limitations of the standardization process (e.g. lack of transparency, resource constraints, stakeholder misalignment)
• Time pressure created by the EU AI Act timeline
• Specific concerns around foundational models and general-purpose AI systems
• Calls for public oversight, more inclusive governance, and stronger institutional checks
Abstract Summary
The article explains the technical standardization framework underlying the EU AI Act, analyzing the strengths and weaknesses of using harmonized standards as regulatory tools. It provides an in-depth critique of the standardization system’s capacity to deliver on the expectations set by the legislation, especially given tight deadlines and the global nature of AI innovation. It also proposes improvements to the standardization governance to ensure legal legitimacy and democratic control.
Document Contents
1. Introduction
2. The Role of Technical Standards in the EU Legal Framework
3. Challenges in AI Standardization
4. Standardization under the AI Act
5. Foundational Models and General-Purpose AI
6. Institutional Reforms and Recommendations
Why it matters?
Harmonized standards are supposed to make the AI Act work in practice. If the system developing them lacks transparency or moves too slowly, the law risks being toothless or unfair. This paper pinpoints where those risks lie and how to address them before implementation kicks off.
What’s Missing?
While the authors clearly explain how technical standards function in law and policy, the piece could benefit from concrete examples of how similar challenges played out in other tech sectors like cybersecurity or medical devices. It also underrepresents the perspective of SME developers or startups, who may face the steepest compliance burdens. The discussion on global standard fragmentation could also go deeper—especially around the potential conflict between EU and U.S./Chinese standards for general-purpose AI.
Best For:
Anyone working on AI compliance in Europe—lawyers, policymakers, and especially technical committee members in standards bodies. Also a solid read for researchers tracking the link between law and technical governance.
Source Details:
European AI Standards: Technical Standardization and Implementation Challenges under the EU AI Act, authored by Nikolas Guggenberger, Eva Schlieck, and Lennart Kroll. Guggenberger is a law professor at the University of Münster and founding director of the Center for Law and Digitalization, known for work on platform governance and tech regulation. Schlieck is a research fellow with expertise in administrative law and EU policy. Kroll is a researcher focused on AI regulation and public law. The paper was published as part of interdisciplinary legal research on European AI policy and provides both legal and technical context for standardization under the EU AI Act.
