Commissioner, Riverside County Flood Control & Water conversation District

A successful first Global Dialogue on AI Governance must transcend high-level declarations and deliver operationalized, technically grounded frameworks. Success should be measured by three concrete outcomes: First, the establishment of Operationalized Interoperability. The Dialogue must bridge the gap between fragmented localized regulations and global technical realities. Success means formally linking UN governance principles to hard, verifiable international technical standards (such as those developed by ISO and the ACM). This ensures that human rights and safety mandates are not just theoretical, but technically auditable across borders. Second, the launch of a Global AI Capacity Clearinghouse. True equity requires moving beyond baseline digital literacy toward "digital mastery" and infrastructure sovereignty. A successful Dialogue will yield binding, multi-stakeholder commitments that channel elite private-sector compute resources, open-source models, and top-tier technical expertise directly to the Global South, actively closing the digital divide. Third, the creation of a UN-Frontier Partnership Pact. Policy cannot outpace technology if the architects of that technology are absent. The Dialogue will only be influential if it systematically integrates the "top one percent" of frontier labs and technical experts into a continuous advisory loop. Success means transitioning from large-scale plenaries to "Evidence-to-Policy" roundtables, where proposed regulations are rigorously stress-tested against state-of-the-art AI capabilities. Ultimately, the session will be a success if it transitions global AI governance from a theoretical listening tour into an actionable, standards-driven ecosystem that mandates safety while democratizing elite innovation.

• Safe, secure and trustworthy AI
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models
• Interoperability of governance approaches

First, Interoperability of governance approaches is the foundational requirement for a cohesive global ecosystem. Fragmented, localized regulations stifle innovation and create compliance vacuums. Drawing from my work on international standards, global governance must be anchored in verifiable technical frameworks to ensure policies operate seamlessly across borders. Second, AI capacity-building must evolve beyond basic digital literacy into infrastructure sovereignty. True equity requires forging binding mechanisms that channel high-performance computing and elite technical expertise directly to the Global South. Developing nations must be empowered to build and govern AI architectures, not merely consume them. Third, Open-source software, open data, and open AI models are the most effective vehicles for safely democratizing frontier capabilities. Cultivating open innovation as a global commons ensures that state-of-the-art AI tools are accessible for advancing the Sustainable Development Goals, preventing the monopolization of technological progress by a few elite labs. Finally, Transparency, accountability, and human oversight prioritize hard, auditable mechanisms over generic calls for "trust." We must establish standards-driven accountability, where human rights and safety protocols are deeply integrated into the engineering and deployment pipelines of frontier models.

Yes, the current thematic areas largely treat AI as static, generative software, overlooking two critical, cross-cutting frontier issues: the rise of Agentic AI and the Physical Compute and Energy Nexus.First, the rapid transition from generative models to Agentic AI-systems capable of autonomous planning, tool use, and executing multi-step actions across digital ecosystems-represents a massive paradigm shift. We are moving from AI as an "advisor" to AI as an "actor." Current governance frameworks are entirely unprepared for autonomous economic agents that can transact, write code, or interact with critical APIs. We urgently need cross-cutting frameworks to address machine liability, agentic identity verification, and boundary controls for autonomous execution.Second, the Physical Compute and Energy Nexus is the invisible chokepoint of global AI governance. The current dialogue heavily emphasizes the digital and social layers of AI but largely ignores the physical realities of its infrastructure. Frontier AI development is fundamentally constrained by access to advanced semiconductors and the massive energy grids required to power hyperscale data centers.Governance must address this physical layer. True "capacity building" and equitable access-the core of the UN's mandate-cannot be achieved without addressing the sovereign distribution of hardware and the environmental impact of AI's energy consumption. By integrating these physical and agentic dimensions, the Global Dialogue can ensure its frameworks regulate the actual frontier of AI, rather than its past.

Operating at the epicenter of AI innovation in California and within the U.S. enterprise sector, the most significant governance gap we face is the fragmentation of regulatory frameworks. The primary challenge is a lack of interoperability. As the Founder of Zenolabs AI and an advisor on international standards, I see firsthand how conflicting state-level regulations, federal directives, and international AI acts create a crippling compliance patchwork. This fragmentation forces enterprise AI builders to spend critical resources on multi-jurisdictional legal engineering rather than advancing technical safety. A related challenge is the concentration of power. The gap in equitable capacity-building means compute and frontier capabilities are increasingly hoarded by a few mega-labs. Relying on these closed systems for transparency and accountability—essentially trusting them to self-police—is unsustainable and stifles broader market competition. The greatest opportunity lies in leveraging open-source models and hard technical standards. By establishing clear, permissive governance for open-source AI, we can prevent monopolization. Open innovation serves as a vital counterweight, allowing agile startups and developing regions to build sovereign, state-of-the-art architectures without gatekeepers. Furthermore, we have the opportunity to solve the accountability gap by anchoring regional compliance in unified global standards, such as those developed by ISO/IEC. By translating vague safety principles into auditable, globally recognized engineering benchmarks, the U.S. AI sector can export not just frontier models, but the operationalized trust required to deploy them safely worldwide.

The Global Dialogue on AI Governance has a singular, vital role to play: it must serve as the operational connective tissue between fragmented regional regulations and the technical realities of frontier AI development. First, it can advance cooperation by shifting the international focus from declaratory principles to operationalized interoperability. Currently, member states are building siloed regulatory frameworks that threaten to fracture the global digital economy. The Dialogue must act as the translation layer, harmonizing these localized policies through hard, globally recognized technical standards—such as those actively being developed by ISO/IEC. By making compliance technically auditable across borders, the Dialogue ensures that AI systems can be deployed globally without compromising sovereign safety mandates. Second, the Dialogue must serve as the primary vehicle for compute and capacity equity. International cooperation is hollow if it only dictates how the Global South should regulate technologies they lack the infrastructure to build. The Dialogue can facilitate a "Global AI Capacity Clearinghouse," creating binding mechanisms to transfer elite private-sector compute resources and open-source models to developing nations, enabling true digital sovereignty. Finally, it must establish a UN-Frontier Partnership Pact. The Dialogue is uniquely positioned to convene the "top one percent" of technical architects. By institutionalizing "Evidence-to-Policy" roundtables, the Dialogue can ensure that international agreements are rigorously stress-tested against state-of-the-art capabilities, guaranteeing that global governance remains tethered to elite technical realities rather than outdated assumptions.

The Global Dialogue must not reinvent the wheel; rather, it should act as the universal integration layer for existing, siloed governance mechanisms. It should explicitly build upon three key pillars: First, Global Technical Standards Bodies. The Dialogue must tightly integrate with the foundational work of ISO/IEC (specifically JTC 1/SC 42) and the ACM. While these bodies produce the rigorous, auditable engineering standards necessary for safe AI, their adoption remains voluntary and fragmented. The Dialogue's added value is to establish Operationalized Interoperability—formally recognizing these technical standards as the de facto baseline for cross-border regulatory compliance and human rights protections. Second, Civic and Policy Frameworks. Initiatives like The Aspen Institute's civic AI programming and the OECD AI Principles have laid the groundwork for democratic, rights-based AI. However, these mechanisms often disproportionately reflect the Global North. The Dialogue's unique added value is scaling these frameworks globally by pairing them with a Global AI Capacity Clearinghouse, ensuring developing nations receive not just policy guidance, but the compute infrastructure and open-source models required to implement it. Third, Frontier Safety Institutes. The Dialogue must connect directly with the emerging network of national AI Safety Institutes (e.g., US, UK) and the Frontier Model Forum. Currently, frontier capabilities and safety insights are concentrated among a few sovereign governments and mega-labs. Through a UN-Frontier Partnership Pact, the Dialogue can ensure that the safety benchmarks developed within these closed-door entities are transparently integrated into global policy and stress-tested by the broader international community. Ultimately, the Dialogue's highest value is convergence: forging a single, actionable global architecture out of disparate technical, civic, and frontier initiatives.

First, Frontier Labs and the Private Sector must contribute more than statements. They must commit computational resources, open-source model access, and empirical safety data to initiatives like a Global AI Capacity Clearinghouse, actively closing the global digital divide. Second, the Technical Community and Standards Bodies (such as ISO/IEC) must provide the hard engineering metrics. Their role is to translate Member State policy objectives into auditable, globally recognized benchmarks, ensuring technical interoperability across jurisdictions. Third, Civil Society and Academia should transition from passive observers to active accountability trackers. Their contribution is to continuously stress-test both frontier models and proposed governance frameworks against human rights mandates and localized realities. Regarding format and structure, large-scale plenary readings are insufficient for regulating exponential technologies. The July Dialogue must pivot to an agile, engineering-focused structure: "Evidence-to-Policy" Roundtables: The Dialogue should be structured around working groups that pair Member State policymakers directly with the "top one percent" of technical architects. This ensures regulations are grounded in elite technical reality rather than outdated assumptions. Technical Stress-Test Segments: The agenda must include dedicated, closed-door sessions where proposed UN frameworks are rigorously stress-tested against the latest advancements, such as Agentic AI and autonomous systems. We must ensure our policies do not expire the moment they are printed. Ultimately, the structure must reflect the urgency of the technology, transforming the UN from a forum for debate into a high-impact operational bridge.

The most glaring omission in global AI governance is not a lack of geographic representation at the microphone, but a lack of infrastructure sovereignty in the Global South. Developing nations are currently relegated to the role of AI consumers and regulatory subjects, rather than AI architects. Furthermore, the independent open-source developer community is critically underrepresented. While frontier mega-labs often dictate the policy discourse, the open-source ecosystem—which serves as the primary engine for democratizing AI capabilities globally—is frequently sidelined in high-level institutional forums. Inclusion must be defined by structural empowerment, not passive observation. To meaningfully include the Global South, the Dialogue must operationalize Compute Equity. Inclusion requires establishing mechanisms to transfer high-performance computing capacity and elite technical expertise to these regions, enabling them to build and govern sovereign AI models. To include the open-source community and independent technical implementers, the UN must shift from declaratory diplomacy to joint engineering. We must elevate technical standards bodies (such as ISO/IEC) as primary governance vehicles and integrate the architects actually building these systems into the policymaking loop. By utilizing "Evidence-to-Policy" roundtables, we can ensure that the developers responsible for democratizing AI are directly involved in stress-testing the frameworks that will govern it. True inclusion requires giving underrepresented voices the tools to build the frontier, not just the opportunity to comment on it.

First, I recommend "Evidence-to-Policy" Red-Teaming Sessions. Instead of reading prepared statements, Member States and the Independent Scientific Panel should engage in live policy "red-teaming." Elite technical architects from the private sector and open-source communities must be brought in to rigorously stress-test proposed governance frameworks against state-of-the-art capabilities, such as autonomous Agentic AI. This ensures regulations are resilient against actual technical trajectories, not just theoretical risks. Second, the Dialogue should host Standards Translation Sprints. We must actively bridge the gap between diplomats and engineers. These focused, working-group sessions would pair policymakers directly with representatives from technical bodies (like ISO/IEC). The sole objective of these sprints would be to take a high-level UN mandate—such as "human oversight"—and immediately translate it into a hard, auditable engineering benchmark. Third, we must operationalize capacity-building through Infrastructure Masterclasses. This means moving beyond digital literacy presentations to live, structural engagements. Frontier labs and infrastructure providers should actively blueprint the transfer of high-performance computing and sovereign model architecture directly with representatives from the Global South. By replacing rhetorical debates with engineering-focused, multi-stakeholder sprints, the Dialogue will transition from a traditional convening into an authoritative, operational bridge

First, Global Technical Standards (e.g., ISO/IEC 42001). The ISO/IEC 42001 AI Management System standard is a premier solution for operationalizing interoperability. It provides a certifiable, universally recognized framework that allows organizations to integrate risk management, transparency, and accountability directly into their engineering pipelines. These hard standards bridge the gap between fragmented regional regulations and global deployment, ensuring compliance is technically verifiable. Second, Responsible Scaling Policies (RSPs) and AI Safety Institutes. The practices being tested by frontier labs and national safety institutes (such as the US and UK AISI)-specifically the practice of tying the training and deployment of models to pre-defined technical safety milestones-offer a rigorous blueprint for managing advanced capabilities, including Agentic AI. The Dialogue should work to internationalize these RSPs, transitioning them from isolated corporate or sovereign policies into a globally harmonized framework. Third, Open-Source Infrastructure Platforms and Sovereign Compute. Ecosystems like Hugging Face have accelerated capacity-building more effectively than any policy declaration by treating open-source models as a global commons. However, software without hardware is insufficient. We must look to emerging practices where nations build "sovereign compute clusters" as a template. By operationalizing the transfer of high-performance computing to the Global South alongside open-weight models, we move beyond basic digital literacy and provide the concrete infrastructure necessary for true digital sovereignty. These examples demonstrate that the most effective governance solutions are not purely legislative; they are structural, standards-driven, and technically anchored.