Robert Kerwin C. Billones, Ph.D.

billion insights by rkb

Universities should lead on AI governance, not wait for it

Six principles and eight actions for higher education institutions, from a white paper I co-led for the Asia-Europe Foundation, and why they matter even more after 2026’s governance developments.

By Dr. Robert Kerwin C. Billones ·

Technology Strategy and Digital Policy

AI regulation around the world is fragmented. The European Union has taken a centralized route with its AI Act and classifies AI systems in education as high-risk. China relies on many departmental rules. Other governments and bodies, from the OECD and the United Nations to Singapore, have issued their own frameworks. For universities that teach, research and partner across these systems, the result is overlapping and sometimes inconsistent expectations.

Universities cannot wait for that to settle. Students carry the way they learned to use AI into their workplaces, so how academia adopts AI shapes how society uses it. That is the starting point of the white paper on AI governance I co-led with Raphaël Weuts of KU Leuven, together with contributing authors from Germany, Poland, Hungary, China, Singapore and the Philippines, through the Asia-Europe Foundation’s Higher Education Innovation Laboratory.

Plan for four futures, not one

Nobody knows how fast AI will advance, so the paper avoids betting on a single forecast. It sets out four scenarios, loosely mapped to the AI safety levels used by frontier AI developers:

  1. A new AI winter. Progress slows and AI stays in non-critical uses such as consumer electronics and specific Industry 4.0 applications.
  2. Sustained progress. AI moves into critical domains such as recruitment, financial evaluation, education, healthcare and policing.
  3. Near human-level AI. Much greater autonomy, from self-driving cars to humanoid robots and an intensified Industry 4.0.
  4. AI beyond the human baseline. Systems that create their own subgoals, with risks that some leading experts describe as existential.

For each scenario we identify no-regret moves: actions that are worth taking whichever future arrives. Equal access, privacy and cybersecurity, transparency about when people are dealing with AI, and clear accountability apply in every scenario. Explainability, human oversight, bias control and sustainable energy and water use become essential as AI moves into critical domains.

Six principles to anchor a university’s position

Drawing on the OECD, G20, EU and Chinese frameworks, we propose six principles: transparency, accountability, safety and reliability, fairness and inclusivity, social impact, and flexibility. The last one matters more than it looks. Because AI keeps changing, rules and good practices need regular revision. A university policy written once and filed away will be out of date within a year.

Eight actions for higher education

  1. Help regulators understand social, economic, sovereign and global AI risks.
  2. Join or build networks of universities working on ethical AI governance, such as an Asia-Europe policy centre.
  3. Adopt internal AI policies, strategies and governance mechanisms built on the six principles.
  4. Integrate AI governance, ethics and safety into existing curricula, with courses for lifelong learners.
  5. Offer executive training and certification in AI governance for leaders in government, industry and academia.
  6. Set up AI innovation sandboxes where new applications are tested responsibly with partner communities.
  7. Develop science advisers and science diplomats who can turn research into policy.
  8. Build thought leadership through public awareness of both the benefits and the risks of AI.

What has changed since we wrote it

The white paper came out of the Asia-Europe Foundation’s laboratory in late 2024 and was published in early 2025. Developments since then have made its recommendations more relevant, not less.

  • Global governance has become a scientific and diplomatic process. The United Nations’ Independent International Scientific Panel on AI, made up of 40 experts from every region, published its first report on July 1, 2026. The first Global Dialogue on AI Governance followed in Geneva on July 6–7, 2026, where speakers warned that capabilities are advancing faster than rules and that the AI divide threatens to leave developing countries behind. That is the science-advice and international-cooperation role the paper asks universities to play.
  • Regulation is arriving unevenly. The EU’s Digital Omnibus, in force since late July 2026, moved most high-risk obligations under the AI Act to December 2, 2027. Education remains a high-risk area. For universities this is a window to prepare, not a reprieve.
  • The Philippines now has national direction. In September 2026, the Commission on Higher Education issued CHED Memorandum Order No. 21, series of 2026, setting guidelines for responsible, human-centred AI use in teaching, learning, research and operations, with safeguards for academic integrity, human judgment, data privacy and student welfare. Our recommended internal policies, governance mechanisms and faculty training are exactly how institutions can put these guidelines into practice.
  • AI is becoming more autonomous. AI agents that plan and carry out tasks are moving into campus systems and workplaces. Greater autonomy is the core of the paper’s third scenario, and it raises the stakes for accountability, human oversight and safety testing.
  • The readiness gap is real. In the Digital Education Council’s 2026 global survey, only 29% of students felt their instructors were equipped to guide them on AI, even though 64% of faculty reported taking AI literacy training. Training alone is not enough. It has to be built into curricula, assessment and institutional practice, as the paper recommends.

Read against these developments, the paper’s no-regret moves and eight actions hold up well. Its scenario approach is also more useful now than a single forecast: with the pace of AI uncertain and regulation shifting, institutions need policies that work in any of the four futures and are revised as things change.

From white paper to practice

I come to AI governance as an engineer. I have built AI systems that run in cities and workplaces, and I have seen that governance works best when it is built into how systems are designed, tested and deployed, rather than added at the end. That is why standards matter. As an expert member of ISO/IEC JTC 1/SC 42, the international committee on AI standards, I see frameworks such as ISO/IEC 42001 as a practical way for institutions to turn principles into management systems.

Several of the paper’s actions have since become concrete work. Action 5 is the idea behind the AI Governance for Higher Education Institutions in Asia program I designed for the United Board for Christian Higher Education in Asia. Actions 1 and 7 run through the Frontiers Champions project on AI governance and safety in the Philippines, supported by the Royal Academy of Engineering. Action 2 is what the Asia-Europe for Artificial Intelligence (AE4AI) Network, where I coordinate AI governance work, is trying to do.

Where a university can start this year

  • Form a small AI governance committee that includes academic, IT, legal, data-privacy and student representatives.
  • Adopt a short AI use policy for teaching, research and administration, and set a date to review it.
  • Keep an inventory of the AI tools already in use, and classify them by risk.
  • Train faculty first. Policies fail when the people expected to apply them do not understand the technology.
  • Add AI governance modules to existing programs before creating new degrees.
  • Measure progress: policies adopted, staff trained, courses updated, and incidents reported and resolved.

Universities are custodians of knowledge. In the age of AI, they also need to be active guardians of the principles that keep AI working for people.

Sources on recent developments

More insights