Event Highlights
Hong Kong Global AI Governance Conference 2026: Highlights
Advancing global dialogue on AI governance beyond conventional geopolitical framings
Organiser
The Hong Kong Global AI Governance Conference 2026 (the Conference), held at the University of Hong Kong on 10–11 April 2026, brought together over 400 participants and 38 speakers from academia, policy, and industry to examine how governance frameworks can respond to rapidly evolving AI systems.
Convened by the HKU Musketeers Foundation Institute of Data Science under IDEAS, the Conference positioned Hong Kong as a platform for cross-regional and interdisciplinary dialogue. Discussions moved beyond simplified geopolitical narratives, focusing instead on the institutional, societal, and technological conditions shaping AI governance in practice.
The opening ceremony was officiated by Professor Xiang ZHANG, President and Vice-Chancellor of HKU, and Mr Ka Chai LEONG, benefactor of HKU IDS, Founder of The Musketeers Education and Culture Charitable Foundation, and Honorary University Fellow 2021. In his welcome address, Professor Zhang highlighted the need to plan ahead for the social, legal, political, and philosophical implications of AI, while Mr Leong underscored the importance of initiatives such as IDEAS in advancing discussion on AI governance.
Across keynote dialogues, fireside conversations, and panel discussions, the Conference addressed the relationship between technological development and governance readiness, the role of institutions in shaping responsible adoption, and the need for coordination across sectors and jurisdictions.
Governance Challenges and Shifting Frames
Across sessions, speakers examined the gap between rapidly advancing AI capabilities and the slower evolution of regulatory and institutional responses. Existing governance models, often designed for more stable systems, face pressure when applied to AI technologies that develop and deploy at speed and scale.
Discussions also revisited the common framing of AI as a geopolitical contest. Rather than treating AI governance solely as a state-level competition, speakers highlighted a broader ecosystem involving companies, research communities, infrastructure providers, and deployment environments.
The Role of Institutions
Universities were consistently positioned as institutions that remain relevant in the AI era, but whose role must evolve. Rather than focusing only on knowledge transmission, higher education was discussed in relation to judgement, critical thinking, interdisciplinary dialogue, and the responsible use of AI.
This reflected a broader view that AI governance extends beyond formal regulation. It also requires sustained engagement across education, research, public discourse, and institutional practice.
AI Governance as an Ongoing Process
The Conference underscored that AI governance is not a one-off policy exercise. It requires continued dialogue across disciplines, sectors, and regions as AI systems and their applications continue to evolve.
Rather than presenting definitive answers, the discussions highlighted the need for governance approaches that can be tested, refined, and adapted in response to real-world deployment.
The Conference received coverage in both English and Chinese media. Reports covered the opening ceremony, Professor Xiang ZHANG’s remarks on the opportunities and challenges brought by AI, the importance of integrating humanities and social sciences with technology, and wider discussions on AI governance, education, and China’s role in global governance efforts.
Day 1 (10 April 2026)
Keynote Dialogue & Opening Fireside Chat
AI’s Future in and Beyond China
AI was framed as a societal technology, where outcomes depend less on capability and more on how systems are designed, deployed, and integrated into human contexts.
Key Discussion Points
- AI as augmentation, not replacement
AI was described as a ‘long-term companion’ designed to extend human capability while preserving human judgement and creativity. - Design and deployment shape outcomes
The impact of AI depends primarily on product design, constraints, and real-world use, rather than on technical capability alone. - Redefinition of human skills
As AI takes on more cognitive tasks, human value shifts towards critical thinking, problem framing, and the ability to evaluate and work with AI systems. - AI as emerging infrastructure
AI was discussed as evolving into a foundational layer of society, raising questions around access, cost, and equitable distribution.
Dr Li XU
Chairman of the Board and CEO, SenseTime, China
Prof Yi MA
Moderator
Director, Musketeers Foundation Institute of Data Science
Director, School of Computing and Data Science
Professor, Chair of Artificial Intelligence, HKU
Panel Discussion 1
AI and Education
The discussion examined how AI is reshaping both the purpose of education and the pace at which institutions can adapt to technological change.
Key Discussion Points
- Reconsidering the purpose of education
Education is shifting from knowledge acquisition towards developing critical thinking, problem framing, and the ability to learn and work with AI systems. - Institutional change is structurally slow
Curriculum reform and system-wide transformation operate on long cycles, creating a lag between technological development and educational adaptation. - Application-led development in practice
In some contexts, AI adoption in education is being driven at scale by national demand and policy direction, particularly in applied and professional training domains. - Human judgement remains central
Students must retain the ability to evaluate AI outputs critically, exercise judgement, and take responsibility for decisions informed by AI.
Dr Lynn DAI
General Manager, SenseTime Education, China
Prof Soraj HONGLADAROM
Professor, International Buddhist Studies College, Mahachulalongkornrajavidyalaya University, Thailand
Prof Tien Yin WONG
Vice Provost, Tsinghua University
Senior Vice Chancellor, Tsinghua Medicine, China/Singapore
Prof Kai Ming CHENG
Moderator
Emeritus Professor, Division of Policy, Administration and Social Sciences Education, Faculty of Education, HKU, HKSAR China
Distinguished Fireside Chat
How AI Can and Should Shape Higher Education
The discussion focused on how universities must redefine their role in an AI-enabled environment, shifting from knowledge delivery to the development of human capabilities that AI cannot replace.
Key Discussion Points
- Universities must adapt, not disappear
AI does not displace higher education but requires institutions to evolve beyond knowledge transmission towards broader developmental and societal roles. - Focus on judgement and inquiry
Universities remain critical in cultivating the ability to ask meaningful questions, form judgements, and communicate ideas effectively. - Human capabilities as core differentiator
Higher education must prioritise imagination, creativity, and independent thinking, rather than competing with AI on information retrieval. - AI literacy across disciplines
Students need to develop the ability to use AI tools critically and responsibly, including understanding their limitations and implications.
Prof Tony CHAN
Third President, King Abdullah University of Science and Technology; Former President, Hong Kong University of Science and Technology, USA/HKSAR China
Prof Duncan IVISON
President and Vice-Chancellor, The University of Manchester, UK
Prof Daniel BELL
Moderator
Chair of Political Theory, HKU School of Governance and Policy, HKSAR China
Panel Discussion 2
The Philosophy of AI Governance: Getting the Concepts Right
The panel examined whether existing conceptual frameworks are sufficient for understanding and governing AI systems, highlighting the risks of relying on imprecise or misleading categories.
Key Discussion Points
Conceptual inadequacy of current terminology
Common terms such as “AI”, “assistant”, and “agent” were seen as too vague or anthropomorphic, potentially distorting both understanding and governance.Limits of human and machine analogies
Framing AI as either human-like or purely mechanical imposes misleading assumptions that fail to capture its hybrid and generative nature.Need for governance-oriented concepts
Effective governance requires conceptual frameworks that go beyond technical description, incorporating responsibility, accountability, and societal impact.Acting under conceptual uncertainty
Governance must proceed despite unresolved philosophical questions, relying on provisional frameworks that can evolve over time.
Prof Herman CAPPELEN
Chair Professor, Department of Philosophy, School of Humanities, HKU, HKSAR China
Prof Mathias RISSE
Director, Carr-Ryan Center for Human Rights, Harvard Kennedy School, USA
Prof Rachel STERKEN
Chair, Department of Philosophy and Associate Dean, Faculty of Arts, HKU, HKSAR China
Prof Jiji ZHANG
Professor, Department of Philosophy, Chinese University of Hong Kong, HKSAR China
Prof Barry SMITH
Moderator
Professor and Director, Institute of Philosophy, School of Advanced Study, University of London, UK
Panel Discussion 3
AI, Law & Regulation
The panel examined how existing legal systems are adapting to AI, focusing on the gap between rapidly evolving technologies and slower regulatory processes.
Key Discussion Points
- Structural tension with existing legal frameworks
Traditional legal models, built on clear causation and identifiable actors, struggle to address the distributed and complex nature of AI systems. - Regulation is inherently reactive
Regulatory approaches tend to respond to known harms rather than anticipate emerging risks, limiting their effectiveness in fast-moving environments. - Trade-offs between protection and innovation
Frameworks such as the EU AI Act and GDPR illustrate the balance between safeguarding rights and imposing compliance burdens that may affect innovation, particularly for smaller firms. - Fragmentation and implementation constraints
Divergent approaches across jurisdictions, combined with practical challenges in compliance and enforcement, create complexity for real-world application.
Prof Pushan DUTT
Professor, INSEAD, Singapore
Prof Mathias RISSE
Director, Carr-Ryan Center for Human Rights, Harvard Kennedy School, USA
Prof Haochen SUN
Professor, Faculty of Law, HKU, HKSAR China
Prof Angela ZHANG
Professor, University of Southern California, USA
Prof Boris BABIC
Moderator
HKU-100 Associate Professor, HKU Musketeers Foundation Institute of Data Science, Department of Philosophy, and Faculty of Law (by courtesy), HKU, HKSAR China
Day 2 (11 April 2026)
Distinguished Fireside Chat
Making Sense of China-US Digital and Tech Governance
The discussion examined AI governance through a geopolitical lens, while challenging simplified narratives and highlighting the complexity of cross-border technological systems.
Key Discussion Points
- Beyond a binary China–US framing
While China and the United States remain central actors, AI development was framed as a broader ecosystem involving multiple countries, institutions, and technological actors. - Governance extends beyond frontier models
AI governance should not be limited to controlling advanced systems, but must address deployment across social, industrial, and institutional contexts. - Divergent regulatory logics and interpretations
China’s approach to AI governance was presented as more complex than commonly portrayed, involving both risk management and support for innovation rather than censorship alone. - Hong Kong as a platform for dialogue
Hong Kong was positioned as a bridging venue capable of facilitating exchange across different systems, regulatory traditions, and policy perspectives.
Mr Kaiser KUO
Writer, technology commentator
Host and Co‑Founder, Sinica Podcast, USA/China
Prof Lan XUE
Dean, Schwarzman College, Tsinghua University, China
Prof Mark WU
Henry L. Stimson Professor, Harvard Law School
Co-Director, Berkman Klein Center for Internet and Society, Harvard University, USA
Prof Brian WONG
Moderator
HKU-100 Assistant Professor, Department of Philosophy
Fellow, Centre on Contemporary China and the World, HKU, HKSAR China
Distinguished Global Conversation
When Nudge Meets AI: The Future of Technology and Human Autonomy
The discussion explored how AI intersects with human decision-making, highlighting both its potential to improve judgement and the risks it poses to autonomy and agency.
Key Discussion Points
- AI can reduce noise and certain biases
AI systems may improve consistency in decision-making by reducing mood-driven variability and some predictable cognitive errors, depending on data quality and design. - Distinguishing influence from manipulation
A key concern is whether AI supports better decision-making or manipulates users by bypassing reflective judgement, raising fundamental questions about autonomy. - Dual role in shaping behaviour
AI can both introduce new forms of manipulation and help users resist them, for example by clarifying preferences or filtering misleading information. - Dependence and distributional concerns
Increased reliance on AI may weaken human judgement over time, while benefits from AI-driven efficiency raise questions about unequal distribution, labour impact, and who ultimately gains.
Prof Cass SUNSTEIN
Professor, Harvard Law School, USA
Prof Brian WONG
Moderator
HKU-100 Assistant Professor, Department of Philosophy
Fellow, Centre on Contemporary China and the World, HKU, HKSAR China
Panel Discussion 4
Values and Institutions for AI Governance
The panel examined which values should guide AI governance and how these can be translated into effective institutional mechanisms.
Key Discussion Points
- Inclusion as a core governance principle
AI governance should address unequal access to infrastructure, education, and opportunity, rather than assuming benefits will be distributed automatically. - Impact on the least advantaged as a benchmark
Technological progress should be evaluated by its effects on poorer and excluded populations, not solely by capability or efficiency gains. - Governance extends beyond the state
Effective governance requires participation from a wider set of institutions, including academia, civil society, media, and international organisations, particularly in contexts of low public trust. - From values to implementation
Normative principles must be translated into practical institutional mechanisms, recognising that policymakers operate under uncertainty and must act without complete information.
Ms Bolor BATTSENGEL
Founder and Chief Executive Officer, AI Academy Asia
Former Vice Minister of Digital Development, UK/MongoliaFounder and Chief Executive Officer, AI Academy Asia
Former Vice Minister of Digital Development, UK/Mongolia
Dr Julian HUPPERT
Founding Director, Intellectual Forum, Jesus College, Cambridge
Former MP for Cambridge, UK
Prof Barry SMITH
Professor and Director, Institute of Philosophy, School of Advanced Study, University of London, UK
Prof Anil GABA
ORPAR Chaired Professor and Academic Director, Centre on Decision Making and Risk Analysis, INSEAD, Singapore
Dr Anoop SINGH
Distinguished Fellow, NITI Aayog and Centre for Social and Economic Progress
Former Managing Director, Asia Pacific Department of International Monetary Fund, India
Prof Daniel BELL
Moderator
Professor and Chair of Political Theory, HKU School of Governance and Policy, HKSAR China
Panel Discussion 5
On the Geopolitics and Global Governance of AI
The panel examined what global AI governance requires in a context of uneven capabilities, diverse political systems, and differing institutional capacities.
Key Discussion Points
- Beyond a two-player geopolitical narrative
Global AI governance cannot be reduced to a US–China binary, but involves a broader ecosystem of states, firms, and institutions with varying roles and capabilities. - Inclusion through access and open systems
Open technologies and shared infrastructure were discussed as mechanisms to broaden participation, enabling more countries and institutions to engage in AI development. - Governance requires enforceable mechanisms
Voluntary principles alone are insufficient; effective governance depends on standards, reporting requirements, procurement rules, and oversight mechanisms that shape behaviour in practice. - Diversity, capacity, and inequality as constraints
Differences in language, culture, and institutional capacity complicate universal governance models, while labour impacts and unequal access remain central concerns.
Prof George CHEN
Partner & Co-Chair, Digital Practice, The Asia Group
Former Regional Public Policy Director, Meta, HKSAR, China
Dr Hongyu FU
Director, AI Governance Center and the Data Economy Center, Alibaba Research Institute, China
Prof Qian XIAO
Vice Dean, Institute for AI International Governance
Deputy Director, Centre for International Security and Strategy, Tsinghua University, China
Dr Genie Sugene GAN
Global Governor, Global Council for Responsible AI
Co-Chair of CSA Alliance
Chair for Cybersecurity, SGTech
Dr Yonghua LIN
Vice President and Chief Engineer, Beijing Academy of Artificial Intelligence, Founder, IEEE WIE, China
Prof Brian WONG
Moderator
HKU-100 Assistant Professor, Department of Philosophy
Fellow, Centre on Contemporary China and the World, HKU, HKSAR China
Concluding Key Dialogue
Are We Educating Our Next Generations to be AI-Agile and -Resilient?
The closing dialogue revisited the role of universities and expert institutions, focusing on how education must evolve in response to AI’s expanding cognitive capabilities.
Key Discussion Points
- Universities remain essential in an AI era
Higher education continues to play a critical role in developing individuals capable of judgement, communication, and participation in society, even as AI advances. - Shift from content delivery to judgement
Educational priorities are moving towards cultivating the ability to ask questions, navigate ambiguity, and critically evaluate AI-generated outputs. - AI literacy as a core capability
Curricula need to incorporate AI literacy across disciplines, including technical understanding, ethical considerations, and responsible use of AI systems. - Lifelong and modular learning models
Education may shift from a one-time degree model to ongoing, modular engagement, allowing individuals to update skills and knowledge as technologies evolve.
Prof Anil GABA
ORPAR Chaired Professor and Academic Director, Centre on Decision Making and Risk Analysis, INSEAD, Singapore
Prof Jay SIEGEL
Vice-President and Pro-Vice-Chancellor (Teaching and Learning), HKU, HKSAR China
Prof Herman CAPPELEN
Moderator
Chair Professor, Department of Philosophy, School of Humanities, HKU, HKSAR China
Closing Keynote
AI Governance: Lessons from Health Care
The keynote used healthcare as a case study to argue that effective AI governance must be sector-specific, system-wide, and grounded in real institutional contexts.
Key Discussion Points
- Sector-specific governance is necessary
Governance approaches must reflect the distinct risks, incentives, and institutional structures of each domain, rather than applying uniform cross-sector models. - System-wide integration over isolated interventions
Effective governance requires understanding how multiple actors—developers, regulators, institutions, and users—interact within a system, rather than focusing on individual components. - Operational systems carry ethical and distributional impact
Routine applications, such as scheduling or administrative processes, can significantly influence outcomes, shaping access, allocation, and fairness in practice. - Trust depends on evidence, not explanation alone
Reliance on explainability is insufficient; trust in AI systems should be grounded in demonstrated performance, validation, and real-world outcomes.
Prof Glenn COHEN
Deputy Dean, Harvard Law School; Faculty Co-Director, Petrie-Flom Center, Harvard Law School, USA
Prof Boris BABIC
Moderator
HKU-100 Associate Professor, HKU Musketeers Foundation Institute of Data Science, Department of Philosophy, and Faculty of Law (by courtesy), HKU, HKSAR China
Conference Album
