The Meta-Layer Initiative

A successful Global Dialogue would move beyond high-level principles toward shared, actionable infrastructure pathways for AI governance. First, success would include alignment on practical governance layers, especially at the point of human-AI interaction. As AI systems increasingly operate across borders and platforms, governance must become more contextual, real-time, and user-facing, rather than relying solely on platform-level or national regulation. Second, the Dialogue should catalyze interoperable trust frameworks, including standards for provenance, identity, and accountability. These should be portable across systems and jurisdictions, enabling collaboration without requiring centralized control. Third, it should produce pilotable governance mechanisms, such as: - transparent AI labeling and interaction signals - user-controlled consent and data-sharing systems - community-informed oversight models Fourth, success would include a commitment to global inclusion, ensuring that developing countries and underrepresented communities can shape and implement governance systems, not just adopt them. Finally, the Dialogue should establish a continuity mechanism, such as working groups or implementation tracks, to translate discussion into iterative deployment and evaluation. In sum, success is not only consensus on values, but the emergence of shared, testable governance infrastructure that can evolve alongside AI systems and support trustworthy, human-centered digital environments.

• Safe, secure and trustworthy AI
• Interoperability of governance approaches
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

These priorities reflect the need to shift AI governance from abstract principles to operational, globally coherent systems. Safe, secure and trustworthy AI is foundational, but must be implemented in ways that are visible and meaningful to users. Trust cannot rely solely on backend safeguards; it must be experienced at the interface level, where people interact with AI systems. Transparency, accountability, and human oversight are essential to maintaining agency in AI-mediated environments. This includes not only explainability, but also clear signaling of AI presence, behavior, and intent, enabling users to make informed decisions in real time. Interoperability of governance approaches is critical in a fragmented digital ecosystem. Without interoperable standards for identity, provenance, and trust signals, governance efforts risk becoming siloed and ineffective. Interoperability enables coordination across jurisdictions while respecting diversity in implementation. Protection and promotion of human rights ensures that governance frameworks remain grounded in dignity, autonomy, and equity. This is especially important as AI systems increasingly shape access to information, economic opportunity, and civic participation. Together, these areas support a model of governance that is distributed, user-centered, and adaptable, capable of responding to the dynamic and cross-border nature of AI systems.

A key emerging issue not fully captured in the listed themes is interface-level governance, or the need to embed governance mechanisms directly into the environments where humans interact with AI. Current approaches focus heavily on model development, data governance, and platform regulation. However, many risks and impacts of AI arise at the point of interaction such as when users encounter AI-generated content, engage with agents, or make decisions based on algorithmic outputs. This suggests the need for a new layer of governance that operates above individual platforms, enabling: - real-time contextual information (e.g., provenance, verification, competing perspectives) - user-controlled consent and data-sharing mechanisms - visible indicators of AI involvement and system behavior - community-driven oversight and feedback loops Such an approach complements existing regulatory frameworks by providing adaptive, context-sensitive governance that can function across jurisdictions and technologies. A second related issue is the development of collective intelligence infrastructure. As AI accelerates information production, the challenge is not only accuracy but also shared sensemaking. Systems that support collaborative interpretation, contextualization, and verification of information will be critical to maintaining social cohesion and informed decision-making. Addressing these emerging areas can help ensure that AI governance is not only protective, but also enabling of trustworthy, participatory digital ecosystems.

Governance gaps in AI are already producing real-world impacts across sectors, particularly in information integrity, digital safety, and economic participation. One of the most significant challenges is the fragmentation of trust and accountability mechanisms. AI-generated content, synthetic media, and automated agents are increasingly indistinguishable from human activity, while existing signals such as verification badges or platform moderation are inconsistent and easily manipulated. This creates risks for misinformation, fraud, and erosion of public trust, particularly in high-stakes domains such as health, finance, and elections. A second challenge is the lack of interoperability across governance approaches. Different platforms, jurisdictions, and technical systems apply divergent standards for transparency, identity, and accountability. This fragmentation limits the effectiveness of oversight and creates regulatory blind spots, especially in cross-border digital environments. A third challenge is the gap between policy and lived experience. Many governance mechanisms operate at the institutional or platform level, but users encounter AI directly at the interface level, where signals about provenance, intent, and reliability are often absent or unclear. This reduces users' ability to make informed decisions in real time. At the same time, there are important opportunities. Advances in areas such as content provenance, decentralized identity, and real-time verification offer pathways toward more trustworthy digital ecosystems. There is also growing momentum around human-centered governance models that emphasize transparency, consent, and accountability in practice, not just principle. If aligned and made interoperable, these developments could enable a shift toward more resilient, participatory, and context-aware governance systems, strengthening both innovation and public trust.

The AI Dialogue can play a critical role as a coordination layer for global AI governance, helping to bridge fragmented approaches across jurisdictions, sectors, and technical domains. First, it can serve as a space to align on shared operational principles, translating high-level values such as safety, transparency, and human rights into practical, interoperable mechanisms. This is essential in a landscape where governance efforts are advancing in parallel but often lack coherence. Second, the Dialogue can enable cross-border experimentation and learning, supporting pilot initiatives and regulatory sandboxes that test governance approaches in real-world contexts. By facilitating structured exchange of results and best practices, it can accelerate convergence without imposing uniform solutions. Third, it can help advance interoperability across governance systems, particularly in areas such as identity, provenance, and accountability. This would allow different national and regional frameworks to function cohesively, reducing gaps that can be exploited in global digital environments. Fourth, the Dialogue can strengthen inclusive participation, ensuring that developing countries, civil society, and technical communities are actively involved in shaping governance frameworks. This is critical for legitimacy and long-term sustainability. Finally, the Dialogue can act as a catalyst for new governance layers that operate at the point of interaction, complementing existing regulatory approaches with more adaptive, user-centered mechanisms. In this way, the AI Dialogue can move beyond discussion to become a platform for coordinated action, enabling a more coherent, responsive, and globally inclusive approach to AI governance.

The AI Dialogue should build upon and connect with a range of existing initiatives across governance, standards, and technical infrastructure. Key efforts include: - OECD AI Principles and GPAI, which provide foundational policy frameworks and multilateral coordination - UNESCO Recommendation on the Ethics of AI, which anchors governance in human rights and global inclusion - ISO/IEC and IEEE standards initiatives, which are advancing technical standards for AI systems - Partnership on AI and other multi-stakeholder initiatives, which bring together industry, academia, and civil society - Content provenance efforts (e.g., C2PA) and emerging identity frameworks, which address trust and authenticity - Regional regulatory frameworks such as the EU AI Act, which are shaping enforceable governance models While these initiatives provide critical building blocks, they often operate in parallel silos, with limited interoperability or coordination at the level of real-world deployment. The added value of the AI Dialogue is its ability to function as a convening and integration layer, connecting these efforts into a more coherent ecosystem. Specifically, it can: - Facilitate interoperability across standards, policies, and technical systems - Support alignment between policy frameworks and user-facing implementation - Enable cross-sector collaboration between governments, developers, and communities - Promote practical governance tools that can operate across platforms and jurisdictions By linking existing initiatives and focusing on implementation pathways, the AI Dialogue can help transform a fragmented landscape into a more coordinated, actionable, and globally inclusive governance architecture.

Different stakeholders can contribute most effectively if the AI Dialogue is structured as a multi-layered, participatory process rather than a single event. Governments can provide policy frameworks and regulatory insights; private sector actors can contribute technical expertise and implementation pathways; civil society can represent societal impacts and rights-based perspectives; and academia can support evidence-based analysis. Critically, these contributions should not occur in isolation but in structured interaction formats that enable dialogue across sectors. To support this, the Dialogue could adopt a three-part structure: - Thematic Working Tracks aligned with priority areas (e.g., trust, interoperability, human rights), enabling focused discussion and outputs - Cross-cutting Integration Sessions to synthesize insights across tracks and identify shared principles and gaps - Implementation Pathways, including pilot proposals, standards alignment, or policy recommendations In addition, the Dialogue should incorporate continuous participation mechanisms, such as digital collaboration platforms, to allow contributions before, during, and after formal sessions. A key recommendation is to prioritize practical outputs, such as: - interoperable governance frameworks - shared terminology and standards - pilot initiatives or testbeds This structure would enable stakeholders not only to share perspectives, but to co-develop actionable governance approaches that can evolve over time.

Several critical voices remain underrepresented in global AI governance discussions. First, Global South stakeholders, including policymakers, technologists, and communities, are often underrepresented despite being disproportionately affected by AI deployment. Their inclusion is essential to ensure governance frameworks are equitable and contextually relevant. Second, youth and future generations are rarely meaningfully included, even though they will experience the long-term impacts of AI systems. Mechanisms for youth participation should go beyond symbolic inclusion toward decision-shaping roles. Third, workers and affected communities across sectors such as education, healthcare, creative industries, and informal economies are often excluded from governance conversations. Their lived experience is critical for understanding real-world impacts. Fourth, non-technical and interdisciplinary perspectives, including cultural, linguistic, and indigenous knowledge systems, are underrepresented. These perspectives are essential for shaping AI systems that respect diverse values and ways of knowing. To address these gaps, the Dialogue could: - Provide accessible participation formats, including multilingual engagement and low-bandwidth options - Support regional and community-led consultations that feed into the global process - Create dedicated representation pathways, such as advisory groups or rotating seats - Offer capacity-building resources to enable meaningful participation Inclusion should be treated not only as representation, but as active co-creation, ensuring that diverse perspectives shape both the process and its outcomes.

To foster meaningful and dynamic engagement, the AI Dialogue should incorporate formats that move beyond traditional panels toward interactive, participatory, and outcome-oriented models. One approach is scenario-based simulations, where participants engage with real-world AI governance challenges (e.g., misinformation events, AI system failures, cross-border data disputes). This allows stakeholders to explore trade-offs and test governance responses in a practical setting. Another format is multi-stakeholder design labs, where small, diverse groups co-develop concrete outputs such as policy prototypes, governance frameworks, or technical standards. These labs can produce tangible results within the Dialogue itself. Real-time polling and feedback systems can also enhance participation, enabling participants to surface consensus, disagreement, and emerging priorities dynamically. Additionally, open annotation or commentary layers on shared documents could allow participants to contribute insights directly to specific ideas, creating a more transparent and collaborative knowledge-building process. The Dialogue could also include regional nodes or hybrid participation hubs, enabling distributed engagement across geographies while maintaining global coordination. Finally, establishing ongoing digital collaboration spaces would allow engagement to continue beyond the event, supporting iterative development and sustained cooperation. These formats help shift the Dialogue from passive exchange to active co-creation, enabling stakeholders to jointly develop governance approaches that are practical, inclusive, and adaptable.

Several existing policies, standards, and technical approaches offer promising pathways for effective AI governance. At the policy level, the EU AI Act provides a risk-based regulatory framework that classifies AI systems by impact and imposes obligations accordingly. Similarly, the UNESCO Recommendation on the Ethics of AI establishes a global, human rights-based foundation for governance, emphasizing transparency, accountability, and inclusion. In the standards domain, initiatives such as ISO/IEC AI standards and the NIST AI Risk Management Framework offer practical tools for assessing and managing AI risks across sectors. These frameworks help translate principles into operational practices for developers and organizations. Technical solutions are also emerging. Content provenance standards, such as those developed by the Coalition for Content Provenance and Authenticity (C2PA), enable verification of the origin and history of digital content, helping address misinformation and synthetic media risks. Model documentation practices, including model cards and system cards, improve transparency by providing structured information about AI systems' capabilities and limitations. In parallel, multi-stakeholder initiatives like the Partnership on AI and the Global Partnership on AI (GPAI) foster collaboration across governments, industry, academia, and civil society, supporting shared learning and best practices. An important emerging approach is the development of user-facing governance mechanisms, such as real-time AI labeling, consent-based data controls, and contextual information layers that provide users with relevant signals about content and system behavior at the point of interaction. Together, these examples illustrate a shift toward governance that is risk-based, interoperable, and increasingly embedded into technical systems, offering concrete pathways to address the complex challenges posed by AI.