What a Shanghai university can teach us about AI governance

When I first arrived at Fudan University this spring to study how one of China's leading institutions is navigating the artificial intelligence revolution in higher education, I was struck by how familiar the challenges sounded.

The debates I listened to in Shanghai — How do we preserve academic integrity? How do we evaluate learning when students can generate essays in seconds? How do we train faculty who feel overwhelmed by the pace of change? — were the same ones echoing through lecture halls in Budapest, Berlin and Paris.

At Fudan, I met the team that built and runs this academic system, and I studied their governance documents in detail. The difference lies not in the problems they face, but in how they have chosen to solve them.

While European universities, including my own at Miskolc, have been exploring creative responses to AI — from structured transparency declarations to new competency-based assessment models — Fudan has done something we have not yet achieved: it has woven these scattered innovations into a single, coherent governance ecosystem. That systemic integration, detailed in the university's recently released guidelines and in my conversations with the people who operate it daily, is what makes the Fudan model worth studying.

The first structural insight is institutional. Fudan created a dedicated AI Education Innovation Center — not a symbolic committee, but an operational body with resources, authority and a clear mandate to connect technological change with daily pedagogical practice.

This center maintains a unified university platform, runs tiered faculty training across three levels (acquire, apply, advance), and continuously updates guidelines as the technology evolves.

In Hungary, we have seen remarkable initiative from individual faculties — ELTE's comprehensive student AI code, BME's problem-solving-focused approach, the University of Miskolc's own framework balancing central principles with local autonomy. What we lack is the institutional anchor that ties these innovations together into a living, adaptive system.

The second insight concerns evaluation. Like many European universities, Fudan initially grappled with the dilemma of how to assess student work in an age of generative AI. Rather than relying on unreliable detection software — a path many of us experimented with in the early days — Fudan shifted to a model of structured transparency.

Students submit declarations with every assignment, documenting which tools they used, how they used them and what human intellectual contribution they added. The university then evaluates AI literacy itself as an academic skill across five dimensions: tool selection, prompt precision, synergistic originality, critical verification and ethical compliance.

This transforms AI use from a hidden activity into a visible, measurable competency. For European institutions that have been independently developing similar ideas, the Fudan framework offers a ready-made architecture.

The third insight is pedagogical. Fudan explicitly encourages what it calls "cognitive load redistribution" — delegating lower-order tasks like information gathering and preliminary structuring to AI tools, so that students can focus their energy on higher-order thinking: critical analysis, complex problem-solving and creative synthesis.

This is not a radical departure from what innovative educators in Europe are already exploring; it is a principled articulation of a direction many of us have sensed but not yet formalized. In my own teaching at Miskolc, I have seen how this framing can transform student engagement when it is supported by clear institutional guidance.

The fourth insight is technological sovereignty. Fudan runs its AI infrastructure on university-controlled servers, ensuring that sensitive research data, student records and institutional knowledge never leave institutional boundaries. For European universities operating under GDPR and the European Union's new AI Act, this principle resonates deeply. The Pannon University in Hungary, for instance, has developed its own research and innovation chatbot for internal use.

The Fudan model suggests that such initiatives should be scaled up and integrated into a unified institutional strategy rather than remaining isolated experiments.

None of this implies that the Chinese model can be transplanted wholesale.

China's centralized governance, its scale and its technological ecosystem differ fundamentally from Europe's decentralized, multi-layered environment.

Hungarian universities operate under a complex mix of national legislation, EU directives and institutional autonomy.

What travels well are the underlying principles: that AI governance should be ecosystem-based, not merely rule-based; that transparency should be operationalized through structured documentation rather than surveillance; that faculty need systematic, tiered training rather than ad-hoc workshops; and that students should be taught to use AI critically and reflectively, not secretly and anxiously.

The first wave of global reaction to generative AI in higher education was defensive — attempts to preserve the pre-AI university through bans, detectors and paper exams.

That phase is passing. What comes next is the harder work of building something new. Fudan's achievement is not that it invented every element of its system from scratch, but that it assembled them into a coherent, operational whole.

For European universities that have been pioneering their own creative solutions, the Fudan model offers not a replacement for our efforts, but a structural blueprint for scaling them up.

The goal is not to preserve the university as it was before AI, but to evolve a new kind of institution — one where human and machine intelligence operate in genuine synergy. That is a project no single country can complete alone.

The author is an associate professor at the University of Miskolc, Hungary.

The views don't necessarily represent those of China Daily.

If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.