Regulating a moving target? Kjetil Rommetveit delivers lecture at The Norwegian Academy of Science and Letters
CAS PI, Professor Kjetil Rommetveit (University of Bergen) delivered a lecture at The Norwegian Academy of Science and Letters, where he was invited to discuss his research on constitutionalism and artificial intelligence.
Rommetveit’s talk explored how the European Union (EU) is attempting to govern the rapidly evolving digital landscape, and what this means for democracy, fundamental rights, and the future of AI.
Since the introduction of the General Data Protection Regulation (GDPR) in 2018, the European Union has intensified its efforts to regulate digital technologies. The EU AI Act, adopted in 2024, is the world's first comprehensive legal framework for artificial intelligence, creating a risk-based system to regulate AI applications and ensure safety and trustworthy AI in Europe.
Basing his lecture on research from his CAS project, “Acting on AI: Digital constitutionalism in the age of AI”, Rommetveit presented a science and technology studies (STS)–based perspective on these developments. He introduced several approaches to (technological) constitutionalism, exploring how they help illuminate the EU’s strategy for governing artificial intelligence.
Drawing on scholars such as Latour, Winner, and Sejersted, Rommetveit situated these developments within the broader field of digital constitutionalism, a framework that examines how constitutional principles are reshaped in a technologically mediated society.
The EU’s attempts to define AI, from early software-focused proposals in 2021 to the final emphasis on “general-purpose models” in 2024, illustrate the complexity of regulating a moving target. The lecture also addressed the problem that the regulation of the industry, in large part, is strongly influenced by the industry themselves, who contribute to shaping the very regulations designed to regulate them.
A central theme of the lecture was how “European values” are translated into practice through concrete governance tools. Rommetveit pointed to the importance of envisioned futures, risk assessments, and standardisation processes, mechanisms that actively shape how AI systems are developed, deployed, and embedded in society.
The CAS Project: Acting on AI
Rommetveit’s project brings together expertise from science and technology studies, law, human–computer interaction, and political science. Its overarching premise is that AI, both as a technological system and as a political project, reshapes fundamental relationships: between humans and machines, states and citizens, and public institutions and private corporations. These transformations, Rommetveit argues, can and should be critically understood through the lens of digital constitutionalism.
Rommetveit emphasised that the AI Act is only one element of a far more extensive regulatory ecosystem. Over the past decade, the EU has developed a dense network of laws governing digital services, media platforms, data governance, and data sharing.
Together, these initiatives reflect a strategic effort to shape the digital market in line with European values and democratic principles. A central question will then be how these values are translated into practice. constitutionalism.