Various, including being a member of UNESCO Women4Ethical AI

A successful first Global Dialogue on AI Governance would move beyond broad principles toward practical governance approaches for real-world AI systems already operating across economies and societies. One important outcome would be recognizing that AI governance is no longer only about models or data, but increasingly about autonomous and semi-autonomous decision-making systems, including AI agents operating in high-impact sectors such as finance. The Dialogue should help establish algorithmic accountability as a core governance priority, particularly in systems where AI influences access to credit, aid, economic participation, and public services. Success would also mean creating clearer pathways for: 1. Explainable AI in high-impact decisions, 2. Embedded ethics by design, 3. Accountability frameworks for AI agents and autonomous AI systems, 4. Meaningful human oversight mechanisms. Another important outcome would be ensuring that governance discussions include perspectives from emerging markets and the Global South, where AI is often being deployed in environments with existing structural inequalities and capacity gaps. Finally, the Dialogue should create continuity mechanisms beyond annual meetings, including sector-specific follow-up tracks and ongoing multi-stakeholder collaboration in rapidly evolving areas such as financial systems and autonomous decision-making infrastructure.

• Safe, secure and trustworthy AI
• Transparency, accountability, and human oversight
• Interoperability of governance approaches
• Social, economic, ethical, cultural, linguistic and technical implications of AI

My priorities focus on how AI systems are increasingly moving from assistive tools toward autonomous and semi-autonomous decision-making systems operating in real economic environments. Safe, secure and trustworthy AI is essential because AI agents are already influencing outcomes in sectors such as finance, where failures can create systemic economic and social consequences. Transparency, accountability, and human oversight are particularly important in high-impact systems where AI decisions affect access to credit, aid, employment, or economic participation. As decision-making authority is increasingly delegated to AI systems, governance frameworks must ensure accountability does not disappear with that delegation. I also selected the broader social, economic, ethical, and technical implications of AI because AI systems trained on historical data can reinforce or automate structural inequalities, including those affecting underserved populations and women. Finally, interoperability of governance approaches is critical because AI systems, financial systems, and digital infrastructure increasingly operate across borders. Fragmented governance approaches may create gaps in accountability, oversight, and implementation.

One important cross-cutting issue that deserves greater attention is the governance of autonomous and semi-autonomous decision-making systems, particularly AI agents operating in real-world economic environments. Current governance discussions often focus on models, data, infrastructure, and general AI safety, but there is comparatively less focus on systems where AI is actively influencing or executing decisions with real human and economic consequences. This is already emerging in sectors such as finance, where AI systems increasingly shape access to credit, onboarding, fraud detection, aid distribution, and economic participation across jurisdictions. In my recent research on AI governance in financial systems, I describe this emerging challenge as: "delegation without accountability." As decision-making authority is increasingly delegated to AI systems, governance frameworks must still ensure accountability, explainability, auditability, and meaningful human oversight. As AI systems become increasingly agentic and interconnected, governance approaches may also need to evolve beyond oversight of individual models toward accountability across interacting AI ecosystems. In recent research, I introduced the concept of "Know Your Swarm" (KYS) governance frameworks to describe the need for visibility, accountability, coordination, and human oversight across interacting AI agents operating collectively within high-impact systems. There is also a growing need to address how AI systems can reinforce or automate structural inequalities, particularly within high-impact economic systems.

In the financial sector, particularly across emerging markets, the Global South, and the Gulf region, AI governance gaps are becoming increasingly visible as AI systems move from assistive tools toward autonomous and semi-autonomous decision-making systems. AI is already influencing outcomes related to credit access, fraud detection, onboarding, compliance, insurance, humanitarian aid distribution, and financial inclusion. However, governance frameworks often remain fragmented and are not evolving at the same pace as deployment. One significant challenge is the growing gap between the speed of AI-driven decision-making and the ability of institutions, regulators, and governance systems to maintain meaningful accountability and oversight. In my recent research on AI governance in financial systems, I describe this emerging challenge as "delegation without accountability." As decision-making authority is increasingly delegated to AI systems and agents, institutions still need to ensure explainability, auditability, accountability, and human oversight, particularly in high-impact systems affecting economic participation. This is especially important in the Global South, where AI is often being deployed in environments with existing structural inequalities, capacity gaps, informal economies, and uneven access to financial infrastructure. Historical financial datasets may already reflect unequal access to capital, undocumented economic participation, and gender-related disparities. Without intervention, AI systems risk reinforcing or automating these inequalities at scale. At the same time, there are major opportunities. If governed responsibly, AI systems can help expand financial inclusion, improve access to financial services, strengthen fraud prevention, reduce operational barriers, and support more inclusive digital economies across developing regions. There are also encouraging governance developments emerging globally. Some recent regulatory approaches are beginning to treat explainability, board-level accountability, and human oversight as operational requirements rather than abstract ethical principles. This shift is important because responsible AI is increasingly becoming part of economic and institutional infrastructure, not just a technology discussion.

The Global Dialogue on AI Governance can play an important role as a neutral, inclusive, and globally representative platform for aligning governance approaches around rapidly evolving AI systems and their real-world impacts. One of its most valuable contributions would be creating a shared space where governments, industry, academia, civil society, and the technical community can collectively address governance gaps that increasingly transcend national borders. This is particularly important as AI systems become more autonomous, interconnected, and embedded within critical sectors such as finance, public services, healthcare, and digital infrastructure. The Dialogue can also help move global discussions beyond high-level principles toward practical governance coordination around: 1. Accountability 2. Explainability 3. Interoperability 4. Human oversight 5. Governance of AI agents and autonomous systems Importantly, the Dialogue can help ensure that emerging markets and Global South perspectives are meaningfully integrated into global AI governance discussions, rather than governance frameworks being shaped primarily by a small number of technologically advanced economies. Another important role would be fostering greater convergence between fragmented governance approaches emerging across jurisdictions. While regulatory diversity will remain important, increasing interoperability and shared governance principles will become essential as AI systems operate across borders and sectors. Finally, the Dialogue can create continuity mechanisms and long-term cooperation frameworks around emerging governance challenges, particularly in high-impact systems where the pace of deployment is already outstripping governance capacity.

The Global Dialogue on AI Governance should build upon existing initiatives such as the UNESCO Recommendation on the Ethics of Artificial Intelligence, the Global Digital Compact, OECD AI Principles, ITU-led discussions, and emerging regional governance approaches. However, the Dialogue's greatest added value should not simply be coordinating existing frameworks, but creating a more inclusive and globally representative governance process. One of the most important gaps in current AI governance discussions is that many conversations remain concentrated among a relatively small group of technologically advanced economies, large institutions, and major technology companies. The AI Dialogue has an opportunity to broaden participation by creating more continuous and accessible engagement mechanisms for: 1. Emerging markets 2. Global South stakeholders 3. Women and underrepresented groups 4. Sector practitioners working directly with real-world AI deployment challenges In particular, the Dialogue could support: 1. Inclusive online working groups 2. Sector-specific collaboration tracks 3. Cross-regional governance exchanges 4. Ongoing multi-stakeholder participation beyond annual meetings This would help ensure that governance approaches are informed not only by frontier technological development, but also by the realities of deployment in diverse economic, cultural, and institutional contexts. As AI systems increasingly operate across borders and sectors, governance discussions must also become more collaborative, interoperable, and globally inclusive. I would also welcome opportunities to contribute to future multi-stakeholder working groups, consultations, and collaborative initiatives emerging from the Dialogue, particularly in areas related to AI governance in financial systems, algorithmic accountability and governance of autonomous AI systems. The added value of the AI Dialogue is its ability to provide a universal platform where a broader range of voices can help shape the future of AI governance collectively.

Different stakeholders bring different forms of expertise to AI governance: 1. Governments provide policy and regulatory leadership 2. Industry contributes implementation realities 3. Academia and technical communities contribute research and emerging insights 4. Civil society contributes inclusion and human rights perspectives To remain effective, the Dialogue should move beyond one-time annual discussions toward more continuous and collaborative engagement. Useful formats could include: 1. Sector-specific working groups 2. Cross-regional roundtables 3. Virtual consultation tracks 4. Practitioner-focused sessions on real-world deployment challenges It would also be valuable to include more discussions focused on high-impact sectors such as finance, healthcare, and digital public infrastructure, where AI systems and agents are already shaping real outcomes. I would also welcome the opportunity to contribute to future multi-stakeholder working groups and collaborative initiatives emerging from the Dialogue, particularly on algorithmic accountability, AI agents, and governance of autonomous systems.

Many AI governance discussions remain concentrated among a relatively small group of technologically advanced economies, large technology companies, and well-resourced institutions. Perspectives from the Global South, emerging markets, women, smaller innovators, and practitioners working directly with real-world deployment challenges remain underrepresented. This matters because AI systems are increasingly being deployed in environments shaped by structural inequalities, informal economies, uneven digital infrastructure, and differing institutional realities. More inclusive participation could be encouraged through: 1. Accessible virtual participation 2. Multilingual engagement 3. Regional consultation tracks 4. Targeted outreach to underrepresented communities 5. Stronger inclusion of practitioners working across AI, finance, ethics, and development More globally representative participation will help create governance approaches that are both more legitimate and more practically effective.

The AI Dialogue could benefit from more interactive and continuous engagement formats beyond traditional panel discussions. Useful approaches could include: 1. Ongoing thematic working groups 2. Cross-regional governance labs 3. Scenario-based policy simulations 4. Practitioner roundtables 5. Virtual collaboration tracks 6. Sector-specific challenge sessions focused on real-world governance issues It would also be valuable to create more collaborative spaces where policymakers, technical experts, industry practitioners, civil society, and Global South stakeholders can directly exchange perspectives and co-develop governance approaches. As AI systems become increasingly autonomous and interconnected, governance discussions will also need to become more adaptive, interdisciplinary and continuous.

Several emerging governance approaches are beginning to move AI governance from broad ethical principles toward more operational and enforceable frameworks. One important development is the growing focus on explainability, accountability, and human oversight in high-impact systems, particularly in sectors such as finance where AI decisions directly affect economic participation and consumer outcomes. There are also important lessons emerging from financial systems and decentralized digital environments, where autonomous and semi-autonomous AI systems have already begun operating under real economic incentives. In my recent research on AI governance in financial systems, I describe one emerging challenge as: "delegation without accountability" where decision-making authority is increasingly delegated to AI systems without corresponding evolution in governance mechanisms. As AI systems become more agentic and interconnected, governance approaches may also need to evolve beyond oversight of individual models toward accountability across interacting AI ecosystems. In my recent research, I introduced the concept of "Know Your Swarm" (KYS) governance frameworks to describe the need for visibility, accountability, coordination and human oversight across interacting AI agents operating collectively within high-impact systems. Effective AI governance will require governance systems that evolve alongside the increasing autonomy and complexity of AI systems themselves.