Sovereign AI: From open foundation models to real-world impact

As artificial intelligence rapidly reshapes economies and societies, the question is no longer whether to adopt AI, but how to build systems that serve public needs and democratic values. This session brings together experts to address two fundamental challenges: 

The first panel examines why open, transparent, and multilingual AI models matter. Beyond serving as an alternative to closed commercial systems, such models can expand access, promote accountability, and enable innovation anchored in inclusivity. The discussion will highlight the value of openness—not only as a technical choice but as a societal one.

The second panel showcases how private-public collaboration and sovereignty can coexist to deliver real impact. Through a concrete case study, panelists will discuss how privacy-preserving AI infrastructure can transform our lives and advance the common good. 

Together these discussions offer actionable models for building AI that earns trust and creates meaningful society values. 

Welcome speech
- Anna Fontcuberta I Morral, President, EPFL

Podium speakers
- Antoine Bosselut, Professor, EPFL 
- Audrey Herblin-Stoop, SVP Public Affairs et Communication, Mistral AI
- Olivier Michielin, Head of Oncology department, University Hospital Geneva 
- Thomas Clozel, CEO, Owkin
- Alix Marxer, General Counsel, WHO Foundation

Moderation: 
- Daniel Dobos, Research Director, Swisscom
- Melissa Anchisi, Head of AI Communication, EPFL

Contact for information, event registration & media requests
- Melissa Anchisi, Head of Communication AI, melissa.anchisi@epfl.ch, +41 21 693 42 08