How should higher education evolve in an AI-driven world?

Governance before tools

One of the Summit’s strongest messages was that AI maturity starts with proper governance. In the opening keynote, Prof. Louis-David Benyayer and Alara Tascioglu introduced ESCP’s ABC framework for AI-ready universities, highlighting that institutions need governance before tools, disciplined experimentation rather than isolated pilots, and a redesign of teaching around judgment and what remains irreducibly human.

This question of governance ran through several sessions. AI adoption is already widespread in higher education, but governance is often fragmented or disconnected from day-to-day practice. Speakers also pointed to a wider strategic challenge: from the EU AI Act to digital sovereignty, universities are being pushed to decide not only how they use AI, but what they control, what they share and what they risk giving away.

In her keynote, Clara Chappaz, French Ambassador for Digital Affairs, framed AI as a geopolitical issue and argued that digital sovereignty must become a strategic concern for higher education. Later, Pénélope Gittos of Hugging Face showed how open platforms can support teaching, research and experimentation while giving institutions greater visibility into the tools they use.

Assessment and pedagogy are being redesigned

Assessment emerged as one of the Summit’s most contested issues. The Summit highlighted a field in transition, with institutions experimenting, retreating or redesigning in response to generative AI. The key takeaway was that assessment reform should be design-led, not fear-led: the goal is to evaluate the process of thinking, framing and deciding, not just the final output that AI can now generate in seconds.

That logic extends into pedagogy. Several speakers argued that if AI can increasingly solve well-defined problems, education must focus more deliberately on problem framing. Speakers described a shift from teaching students to produce answers towards helping them ask better questions, work with ambiguity and understand when to use — or refuse — AI.

Research integrity and sustainability are no longer side issues

The Summit also highlighted two areas that are often treated as secondary, but cannot remain so: research integrity and sustainability. On the research side, contributors raised concerns around hallucinations, industrialised paper fraud, reviewer fatigue and academic deskilling.

The research dimension was also explored by Sébastien Bubeck of OpenAI, whose keynote examined how advanced AI systems are transforming the research process itself, from literature review to experimentation and discovery.

On sustainability, a closing roundtable featuring Gorgi Krlev, Associate Dean of Sustainability at ESCP, Sandrine Kergroach of the OECD, and Mario Calderini from Politecnico di Milano made a deliberate point: the final measure of AI integration is not what it enables, but what kind of institutions and societies it helps build.