Session
Organizer 1: Yik Chan Chin, Beijing Normal University
Organizer 2: J Amado Espinosa L, Medisist
Organizer 3: Sergio Mayo Macias, Instituto Tecnológico de Aragón
Speaker 1: Dr Yik Chan Chin, Associate Professor, Beijing Normal University, China ; Co-leader of the IGF Policy Network on AI (PNAI)
Speaker 2: Mr. Poncelet Ileleji, CEO - Jokkolabs Banjul /NRI Gambia, Africa
Speaker 3: Mr. Sam Daws, Senior Advisor, Oxford Martin AI Governance Initiative, Oxford University, UK and Director, Multilateral AI
Speaker 4: Dr Xiao Zhang, Deputy Director, China Internet Network Information Center (CNNIC) & Deputy Director of China IGF
Speaker 5: Mr. Mauricio Gibson, Head of International AI Policy, AI Policy Directorate, Department for Science, Innovation and Technology, UK
Speaker 6: Dr. Neha Mishra, Assistant Professor, Geneva Graduate Institute, Switzerland
Dr Chafic Chaya, Technical Community, Western European and Others Group (WEOG)
Mr. Heramb Podar, Civil Society, Asia-Pacific Group
Yik Chan Chin, Civil Society, Asia-Pacific Group
Roundtable
Duration (minutes): 90
Format description: The session will focus on a specific Internet governance topic, the topic is focused but also broad that will require in-depth deliberation and debate by five experts with diverse perspective, gender, region, and stakeholder group , thus 90 minutes roundtable is the best suited format to allow a sufficiently deep and effective deliberation and interaction.
-
Understanding the Interoperability of AI Governance
What is your understanding of interoperability, and what are the most issues that need to be addressed at the global level, and what are the obstacles? -
Addressing Interoperability Issues
How can different actors address interoperability, and how can we balance regional variations with global approaches? -
The Role of the UN in Global AI Governance
What role should the UN play in tackling AI governance?
What will participants gain from attending this session? The participants and attendees will gain new knowledge, insights and understanding of 1) individual country and region’s current AI interoperability efforts 2) The scope of and most important mechanisms of achieving the interoperability of AI Governance from technologic, regulatory and semantic perspectives 3) How to promote the collaborative and inclusive international cooperation to achieve the AI governance's interoperability
Description:
Interoperability is often understood as the ability of different systems to communicate and work seamlessly together. OECD AI principles, recent UN Resolution on AI, and UN HLAB on AI's interim report all stresses on the importance of developing interoperable AI. We need to analyse if and how different regulations/norms/standards/ to govern AI could work together. The 2023 IGF PNAI report defines the interoperability in AI as a framework that brings together of three key aspects: (1) the substantive mechanisms involved in guiding and developing of digital technology; (2) multistakeholder interactions and interconnections, and (3) agreed ways to communicate and cooperate. Recent developments in the EU and ASEAN regions offer valuable insights emphasising transparency, accountability, and defining interoperability requirements that ensure AI systems are comprehensible and operate consistently across different regions. This workshop will address the scope and mechanism of interoperability of AI governance from three aspects: 1.) Identify the most important and pressing regulatory, technological, data and semantic issues of AI governance needed to be addressed at the global level ; 2) define the principles, methodologies and mechanisms should be applied; 3) Promot collaborative and inclusive international cooperation to align methodologies and mechanism that promote innovation and prevent the fragmentation of the governance of AI systems. For instance, potential agreement on the standards for a technical exchange of data, testing, training and deployment etc.
Agenda:
1) Introuction to speakers 10 mins
2) Three policy questions deliberation 40 mins
3) Feedback on delibration 10 mins
4) Q&A 30 mins
Multistakeholder speakers and moderators from Africa, China, Inidia, EU, UK Middle East will discuss above questions from diverse geographic and stakeholders’perspectives.
1) Deliberate and identify the scope and most important mechanisms of achieving the interoperability of AI Governance from technical, regulatory and semantic perspectives; 2) Identify the different AI governance interoperability efforts and problems in speakers’ countries and region; 3) Develop a framework for collaborative response that includes multi-stakeholders (not only governments) such as technical sector, civil society, private sector and other specialists to provide a meaningful platform that tackles the interoperability of AI governance. 4) Policy recommendations and key messages report to the UN and regional/national IGF communities, private AI companies, and other relevant epistemic communities including thinktanks such as Oxford Global Society, Industrial associations such as China’s Artificial Intelligence Industry Alliance etc
Hybrid Format: 1) The workshop has an onsite moderator and online moderator, each is responsible for moderating the onsite and online speakers and attendees. And both moderators will ensure all speakers and attendees no matter online and onsite will have the equal opportunity to speak, raise questions and engage in each session of the workshop. 2)The session will be moderated by the online and onsite moderators. The moderators will invite each speaker to express their views on a set of questions and guide the debate amongst speakers and the audience, moderators will invite questions from the onsite audience and online participants, the question time will last about 30 minutes in order to provide sufficient interactions amongst speakers, audience and online participants. 3)The social media platforms including Youtube, X, Facebook, Weibo and Wechat etc will be used to increase the participation and to carry out online streaming of the session,
Report
The UN acts as a key point of disemmination of information to member states and to build mutual understanding in terms of AI governance. However, it is only as powerful as it's member states are willing to let it be. There are prevailing tensions between multilateralism and multistakeholderism which need to be adressed at a global level. Next steps are to find priorities to focus on a national, regional and international level.
Reduce overlapping agendas, capacity building on a grassroots and governmental level. There needs to be consideration on different domains of interoperability like cultural, linguistic, ethical, technical and data.Establish a global AI policy dialogue in IGF. To maximize IGF's potential for delivering concrete outcomes, long-term sustainability needs to be ensured through increased financial, technical and human resources support.
The panel invited experts from China, the UK, Europe, and Africa to discuss the interoperability of AI governance, aiming to promote global cooperation and dialogue. Through expert presentations, deliberations, and interactions, the roundtable discussed the challenging questions of: 1. Understanding the Interoperability of AI Governance: What is your understanding of interoperability, and what are the most issues that need to be addressed at the global level, and what are the obstacles? 2: Addressing Interoperability Issues: How can different actors address interoperability, and how can we balance regional variations with global approaches; 3. The Role of the UN in Global AI Governance What role should the UN play in tackling AI governance?
Key Discussion Points:
1. Definition of Interoperability: Yik Chan Chin: focused on interoperability in legal, semantic, and technical layers. It involves different initiatives such as laws, regulations, policies, and standards working together across legal, semantic, and technical layers to be more effective and impactful globally. Sam Daws: added sustainability and energy efficiency concerns as interoperability aspects. Neha Mishra: Emphasized the "human layer" of interoperability, which involves cultural and ethical considerations in AI governance.
2. Cultural Interoperability and Its Importance: Sam Daws: stressed the need for "cultural interoperability" to ensure that AI systems respect and integrate diverse cultural values and practices globally. Warned against the risk of cultural fragmentation, such as isolated "cultural AI ecosystems" (e.g., the risk of a bifocation of socially-conservative alliances like BRICS AI, and Western cultural approaches). Advocated for promoting culturally inclusive AI systems as part of global collaboration. Neha Mishra: Advocated for capacity-building that respects the cultural autonomy of developing countries. Suggested a cautious approach to avoid imposing external frameworks that undermine local practices. Xiao Zhang: Highlighted the importance of understanding cultural contexts when developing AI governance systems, contrasting AI governance with internet governance. Argued for AI governance approaches to be culturally aware and sensitive to the risks AI poses, such as militarization and misuse.
3. Global Issues and Obstacles: Yik Chan Chin highlighted global issues of risk categorization, liability in AI systems, data training risks, and alignment of technical standards and obstacles like geopolitical conflicts (e.g., high-end chip restrictions), lack of trust among countries, unequal distribution of AI technology, and maturity of policymaking. Mauricio Gibson: Emphasized the practical challenges in governance, including varied government interests and policy-making maturity in the face of fast technological change. Suggested building trust and clarity in messaging and improving coordination among multilateral domains. Xiao Zhang: Distinguished AI governance from Internet governance, focusing on its higher stakes (e.g., military risks). Stressed the importance of multilateral collaboration for AI safety and resource mobilization.
4. Balance Regional Variations with Global Approaches: Yik Chan Chin: Advocated for multistakeholder and multidisciplinary approaches, with alignment at regional, and global levels. Suggested international safety collaborations and mechanisms like mutual recognition of standards and security certifications. Sam Daws: Proposed early regional collaborations to reduce conflicts and using cross-regional forums for interoperability. Highlighted contributions from organizations like the African Union and Digital Cooperation Organization (DCO). Mauricio Gibson: Discussed the role of governments in convening different stakeholders and understanding issues to funnel them into policy. Highlighted the need for capacity building , particularly in developing regions and clear messaging to different communities. Mentioned the UK's investment in AI development programs and the importance of supporting policy officials and civil servants. Shared examples of regional dialogues (e.g., OECD-African Union workshops) to bridge global and regional approaches. Xiao Zhang: Called for multinational frameworks with multilateral resources to address the unique risks of AI.
5. Role of Governments: Mauricio: Governments play a crucial role in convening different stakeholders and understanding issues to funnel them into action and policy. Capacity building and clear messaging are essential for different communities across the world.
6. United Nations' Role: Yik Chan Chin: Supported the UN as a platform for dialogue, interoperability and trust-building while highlighting the need for IGF and multistakeholder dynamics. Sam Daws: Emphasized the UN’s role in offering leadership, building trust through capacity-building and scientific collaboration. Suggested leveraging existing initiatives (e.g., Sustainable Development Goals) and focusing on areas with prior cooperation (e.g., food security, climate, health). Pointed out that member states currently lack the appetite to grant the UN regulatory or enforcement powers over AI governance. Mauricio Gibson: Advocated for the UN to enhance coordination without duplicating efforts and to use its convening power to facilitate forums and stakeholders from across the world in sharing their best practices.
5. Multistakeholder and Multilateral Engagement: Yik Chan Chin: Multistakeholder Approach: Emphasized the importance of collaboration among diverse stakeholders. Highlighted the need for regional alignment (e.g., within African or Latin American unions) before moving to global interoperability. Multilateral Coordination: Stressed the role of global platforms like the UN in coordinating national, regional, and global efforts without compromising national diversity. Supported the integrating multistakeholder efforts with multilateral initiatives to leverage diverse resources and expertise. Sam Daws: Multistakeholder Engagement: Advocated for cross-regional forums to reduce silos in AI approaches. Highlighted the value of transnational collaboration between political, cultural, and scientific communities. Called for using bodies like the Digital Cooperation Organization and regional organizations for wider dialogue. Multilateral Focus: Suggested early-stage cooperation in drafting national and regional approaches to ensure alignment with global interoperability. Proposed leveraging the UN’s capacity for trust-building and scientific exchange while recognizing its current limits in enforcement. Mauricio Gibson: Multistakeholder Efforts: Highlighted the role of governments in convening and supporting diverse stakeholders. Shared examples of initiatives like the UK-funded AI Standards Hub to foster international collaboration. Multilateral Engagement: Advocated for clarity and coordination among multilateral forums to avoid duplication of efforts. Shared examples of multilateral cooperation, such as OECD-African Union dialogues and initiatives for capacity building in developing regions. Xiao Zhang: Multistakeholder Approach: Emphasized that AI governance, unlike internet governance, requires greater involvement due to its potential risks. Multilateral Focus: Supported the idea that AI governance should rely on multilateral frameworks, led by organizations like the UN, to ensure global alignment and safety. Poncelet Ilelegi: Multistakeholder Engagement: Focused on inclusivity, emphasizing that stakeholders from the Global South need stronger representation in multistakeholder dialogues. Highlighted the importance of AI governance frameworks enabling trust and cooperation among various actors. Multilateral Engagement: Supported the idea of global frameworks like the UN acting as umbrella organizations to bridge regional divides and promote shared goals, such as achieving the SDGs. Neha Mishra: Multistakeholder Engagement: Highlighted the alignment challenges between multistakeholder and multilateral approaches but stressed the importance of collaboration among governments, academia, and the private sector. Called for frameworks that allow for the autonomy of developing countries while promoting inclusivity. Multilateral Engagement: Advocated for modular and incremental approaches to multilateral cooperation to address disparities in technological and governance capacities across regions.
Implementation and Coordination: Yik Chan Chin: Proposed mechanisms like mutual recognition agreements and security certifications to streamline global standards. Advocated for multi-level coordination: Local efforts could align with regional bodies (e.g., African Union, EU). Regional policies can escalate to global frameworks, facilitated by platforms like the UN. Highlighted the importance of reducing unnecessary barriers and costs by aligning domestic and international regulatory efforts. Sam Daws: Stressed the importance of early regional and cross-regional cooperation in drafting policies to avoid conflicts later. Highlighted the need to build frameworks that integrate sustainability and transparency, using examples like Singapore’s AI Verify initiative. Urged the UN to act as a hub for capacity-building efforts while maintaining clarity on its coordination responsibilities to avoid duplication of activities. Mauricio Gibson: Shared examples of actionable initiatives, such as: The UK’s AI Standards Hub, which facilitates international coordination of standards. OECD-African Union workshops that align regional and global principles. Highlighted the need for practical tools and frameworks, such as AI self-assessment tools, to enable real-world adoption of policies. Advocated for avoiding duplication in multilateral forums and ensuring clear messaging about responsibilities and roles. Highlighted resource allocation challenges, particularly for developing countries, and the need for governments to channel resources effectively. Xiao Zhang: Proposed that nations focus on aligning local AI development with broader multilateral objectives, especially given the risks associated with unregulated AI development. Highlighted that global implementation should consider cultural and economic differences. Stressed the importance of multilateral frameworks (e.g., the UN) to guide national efforts while ensuring they respect national sovereignty and cultural diversity. Advocated for leveraging the Internet Governance Forum (IGF) as a platform for coordination and dialogue. Neha Mishra: Warned against imposing universal models that could undermine the autonomy of developing countries in shaping their AI governance. Highlighted the challenge of coordinating multistakeholder and multilateral efforts, advocating for clarity and collaboration across forums. Stressed the importance of avoiding duplication of efforts while ensuring diverse stakeholders have representation in global initiatives.
Q & As:
- Cataldo Dell'Accio (onsite audience): Highlighted the convergence of AI and blockchain, suggesting blockchain could enhance interoperability by providing a common layer of trust.
- Online Audience Questions: Focused on the slow pace of multistakeholder bodies like the UN, the need for efficiency improvements, and balancing AI sovereignty with global interoperability.
- Sam Daws: Supported blockchain for improving accountability and proposed "data embassies" for secure country-specific data storage. Emphasized the dual need for the UN to balance slow, legitimacy-building norm-building with the ability to respond quickly as already existed in the peace and security arena. But the UN would likely still rely on member state monitoring and early warning capabilities in relation to AI threats
- Dr. Yik Chan Chin: Agreed on blockchain's role in security and stressed reducing overlaps in UN agents. Highlighted the need to balance global frameworks with national sovereignty in AI governance.
- Xiao Zhang: Called for leadership engagement in AI governance and suggested prioritizing AI-related issues on the UN agenda while strengthening multistakeholder approaches.
Major future challenges:
- Geopolitical Fragmentation and trust deficits are recurring challenges across speakers.
- Sustainability, digital divides, and capacity gaps require urgent attention.
- A balance between national sovereignty and global interoperability must be carefully managed.
- Highlighted the need for multilateral bodies like the UN to act both swiftly and thoughtfully in responding to emerging AI challenges.
- Cited the rapid evolution of AI as a challenge for policy frameworks to remain relevant.
- Clearer coordination and leadership are needed to address these future challenges effectively.
Final Remarks: The panel emphasized the importance of public interest, inclusivity, and the need for continued efforts to find common points of consensus despite geopolitical and developmental differences.
Read the coverage of workshop by Geneva Internet Platform:
Watch the panel discussion recording at: https://www.youtube.com/watch?v=AsBtfX-hGMc