In February 2025, OFCOM published—at the Federal Council’s request—three major reports outlining possible options.

This resulted in (1) the signing of the Council of Europe Convention on AI (on March 26, 2025), and (2) the decision to adopt a sector-specific approach, in contrast to the horizontal approach taken by the European Union with its AI Act, with a draft expected by the end of 2026.

This choice is in keeping with Swiss tradition: regulate only where necessary.

However, the question remains open as to how to build a coherent framework in an environment marked by (1) significant information asymmetry, and (2) a proliferation of standards, initiatives, and reports at the international level.

It is precisely to contribute to this discussion that I am pleased to participate this afternoon, at the invitation of Yaniv Benhamou (UNIGE), Florent Thouvenin (UNIZH), and Nadja Braun Binder (UNIBA), in a forum bringing together some fifteen experts from various backgrounds, under the auspices of the Digital Law Center – University of Geneva Center.

I am particularly looking forward to these discussions, which are essential for developing a Swiss approach that is pragmatic, proportionate, and compatible with economic realities.

While everyone recognizes the need to regulate the development and deployment of AI tools given the risks (some of which are well-established), the information asymmetry and the proliferation of international discussions and reports on the subject make it difficult to reach a consensus.

I can only welcome the efforts undertaken by my colleagues to establish a think tank on these issues and am already looking forward to the discussions on this topic within a small multidisciplinary group.