BUEM

The success of the first Global Dialogue on AI governance, initiated by the United Nations under resolution 79/325, should be measured not only by the breadth of participation but by tangible and actionable outcomes. First, a key achievement would be the agreement on a set of shared principles for AI governance, reflecting a broad consensus among governments, industry, academia, and civil society. These principles should address safety, transparency, accountability, and the protection of human rights, serving as a reference point for national policies and corporate practices. Second, success would require the establishment of sustainable mechanisms for international cooperation. This includes the creation of thematic working groups, regular consultations, and structured exchanges of best practices. The Dialogue should move beyond general discussions toward continuous collaboration on concrete issues such as risk assessment, data governance, and algorithmic auditing. Third, meaningful inclusion of developing countries is essential. A successful outcome would involve commitments to capacity-building, including access to technology, training, and institutional support, thereby helping to bridge the global digital divide. Fourth, the Dialogue should ensure the integration of scientific expertise into policymaking, including a clearly defined role for an independent international scientific panel on AI and mechanisms for its engagement with governments. Finally, success depends on the adoption of a clear roadmap for future action, including a schedule of meetings, defined mandates for working bodies, and measurable objectives. Without this, the Dialogue risks remaining purely declaratory. In sum, a successful first Global Dialogue would mark the transition from discussion to the establishment of an inclusive, durable, and results-oriented framework for global AI governance.

• Safe, secure and trustworthy AI
• AI capacity-building
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

From the listed thematic areas (as outlined by the United Nations in resolution 79/325), the following four priorities would be the most relevant for urgent action and active engagement: 1. Development of safe, secure, and trustworthy artificial intelligence systems This is a foundational priority, as ensuring the technical reliability and safety of AI systems is essential for minimizing risks and preventing harm. 2. Capacity-building in artificial intelligence Strengthening skills, infrastructure, and institutional capabilities-especially in developing countries-is critical to ensuring inclusive participation and reducing the global digital divide. 3. Protection and promotion of human rights AI systems must be aligned with international human rights standards to prevent discrimination, safeguard freedoms, and ensure ethical deployment. 4. Transparency, accountability, and human oversight These elements are essential for building trust in AI systems, enabling explainability, and ensuring that responsibility for outcomes remains clearly defined. Together, these priorities balance technical safety, social equity, and governance accountability, making them both practical and strategically important areas for engagement.

Yes, while the listed themes under the United Nations resolution 79/325 are comprehensive, several cross-cutting and emerging issues merit additional attention. One important gap is the environmental impact of AI systems. The rapid growth of large-scale models and data centers has significant implications for energy consumption, water use, and carbon emissions. Addressing AI sustainability should be integrated into governance discussions to align with global climate goals. Another emerging issue is the concentration of power in the AI ecosystem. A small number of companies and countries currently control advanced computing resources, data, and frontier models. This raises concerns about market dominance, dependency, and unequal influence over global standards and norms. A third area is information integrity and societal resilience. While risks like misinformation are often mentioned, there is a need for deeper focus on how AI affects democratic processes, public discourse, and trust in institutions, especially in multilingual and culturally diverse contexts. Additionally, security risks at the intersection of AI and cybersecurity require more explicit attention. AI can both strengthen and undermine cybersecurity, enabling more sophisticated attacks such as automated phishing, deepfakes, or system vulnerabilities. Finally, there is a need to address long-term governance challenges, including the pace of technological change relative to regulatory processes, and the question of how to ensure that governance frameworks remain adaptive and forward-looking. In sum, incorporating sustainability, power concentration, information integrity, cybersecurity, and long-term adaptability would strengthen the Global Dialogue and ensure a more holistic approach to AI governance.

From the perspective of Slovakia and the broader Central and Eastern European region, gaps in AI governance—and progress in the selected priority areas—have both significant risks and opportunities. A key challenge is limited capacity and uneven access to AI infrastructure. Compared to larger economies, Slovakia faces constraints in computing resources, investment, and specialized talent. Without targeted capacity-building, this may deepen dependency on external providers and limit domestic innovation. Another challenge relates to regulatory fragmentation and implementation gaps. While the European Union is advancing a comprehensive framework (notably the AI Act), differences in national readiness, institutional capacity, and enforcement may create inconsistencies. This can affect businesses operating across borders and slow adoption of trustworthy AI systems. At the same time, there are important opportunities. Slovakia can benefit from aligning with EU standards to position itself as a trusted environment for AI deployment, particularly in sectors such as manufacturing, automotive, and public administration. Strengthening transparency, accountability, and human oversight can enhance public trust and attract investment. Capacity-building is also a major opportunity. Investments in education, research, and digital infrastructure could enable Slovakia to develop niche expertise, especially in applied AI and industrial use cases. Collaboration within the EU and participation in global initiatives under the United Nations framework can further amplify these efforts. Finally, advancing human rights–based and trustworthy AI offers a strategic advantage. By embedding ethical standards early, Slovakia can avoid reputational risks and ensure that AI adoption supports social cohesion rather than exacerbating inequalities. Overall, the main challenge lies in closing capacity and implementation gaps, while the key opportunity is to leverage regional cooperation and global frameworks to build a competitive and responsible AI ecosystem.

The Global Dialogue on AI, established under the United Nations resolution 79/325, can play a central role in strengthening international cooperation on artificial intelligence governance by acting as an inclusive and continuous multilateral platform. First, it can serve as a bridge between different governance approaches. Countries are currently developing AI regulations at different speeds and with varying priorities. The Dialogue provides a neutral space where these approaches can be compared, aligned where possible, and made more interoperable, reducing fragmentation in global AI governance. Second, it can facilitate the development of shared norms and principles. While it is unlikely to produce binding treaties in the short term, it can help build convergence around core ideas such as safety, transparency, accountability, and respect for human rights. These shared norms can gradually influence national policies and industry standards. Third, the Dialogue can strengthen capacity-building and inclusion, particularly for developing countries. By enabling knowledge-sharing, technical assistance, and access to best practices, it helps ensure that AI governance is not dominated only by technologically advanced states, but reflects global participation and equity. Fourth, it can improve coordination between stakeholders beyond governments, including industry, academia, and civil society. Given the rapid pace of AI development, multi-stakeholder engagement is essential for identifying risks early and developing practical responses. Finally, the Dialogue can support early identification of emerging risks and opportunities, functioning as an anticipatory governance mechanism. Regular discussions and inputs from the independent scientific panel can help the international community respond proactively rather than reactively to technological developments. In sum, the Global Dialogue can act as a coordination hub that fosters trust, reduces fragmentation, and gradually builds a more coherent and inclusive global framework for AI governance.

The Global Dialogue on AI under the United Nations resolution 79/325 should build on, and actively coordinate with, several existing international and regional initiatives in order to avoid duplication and strengthen coherence in global AI governance. First, it should engage closely with the European Union AI regulatory framework, particularly the EU AI Act and related standards work. The EU is currently one of the most advanced regulatory actors, and its experience in risk-based regulation can provide valuable practical lessons. Second, it should connect with OECD AI governance initiatives, especially the OECD AI Principles, which already serve as a widely referenced soft-law framework among developed and developing countries. These principles offer a useful foundation for shared norms. Third, collaboration with the Global Partnership on Artificial Intelligence (GPAI) is important, particularly in areas of technical research, responsible innovation, and multi-stakeholder engagement. GPAI's working groups can provide evidence-based input to the Dialogue. Fourth, the Dialogue should interact with international standards bodies such as ISO/IEC, which are developing technical standards for AI safety, risk management, and interoperability. This ensures that governance discussions translate into implementable technical norms. The unique added value of the Global Dialogue lies in its universality and political legitimacy. Unlike technical or regional initiatives, it provides a truly global forum under the UN umbrella where all Member States—regardless of technological capacity—can participate on equal footing. Its second key advantage is its integrative function: it can connect fragmented initiatives, align norms, and facilitate coherence between policy, technical standards, and ethical frameworks. Finally, its ability to link policy discussions with the independent international scientific panel on AI gives it a strong evidence-based foundation, enabling it to translate scientific insights into global governance discussions. In this way, the Global Dialogue acts not as a competitor to existing initiatives, but as a coordinating hub that enhances their collective impact.

Different stakeholders can contribute meaningfully to the Global Dialogue on AI under the United Nations through clearly defined, structured participation mechanisms that ensure inclusivity, expertise, and balanced representation. Governments should play a central role in setting priorities, sharing national regulatory experiences, and identifying policy gaps. They can contribute through formal plenary sessions, thematic negotiations, and submission of national reports on AI governance practices. Industry and private sector actors are essential for providing technical expertise and real-world insights into AI development and deployment. Their participation should be structured through dedicated industry roundtables, technical working groups, and voluntary commitments on safety, transparency, and responsible innovation. Academia and the scientific community, including the independent international scientific panel on AI, should provide evidence-based assessments of risks, capabilities, and emerging trends. Their input should be systematically integrated through regular scientific briefings and peer-reviewed reports presented to the Dialogue. Civil society organizations play a critical role in representing public interest, human rights concerns, and social impacts. Their engagement should be ensured through open consultations, stakeholder forums, and structured input mechanisms that allow them to influence agenda-setting. Recommended format and structure for the Dialogue should include: * A high-level annual plenary session for strategic political discussion; * Thematic working groups focused on priority areas such as safety, governance, and capacity-building; * A multi-stakeholder advisory forum ensuring continuous input from non-state actors; * A science-policy interface mechanism linking research findings to policy discussions; * And an implementation and review track to monitor progress and share best practices. To be effective, the Dialogue should remain flexible, transparent, and iterative, allowing for both formal negotiation and informal knowledge exchange. This structure would ensure that contributions from all stakeholders are not only heard but systematically integrated into global AI governance processes.

In global discussions on AI governance, including those under the United Nations framework, several important communities remain underrepresented, limiting the inclusiveness and legitimacy of decision-making. First, developing countries and least developed countries are often insufficiently represented, particularly in technical and policy-shaping discussions. Limited access to expertise, funding, and infrastructure constrains their participation. As a result, global AI norms risk reflecting mainly the priorities of technologically advanced states. Second, Indigenous peoples and local communities are frequently absent, despite being directly affected by AI systems in areas such as data governance, cultural representation, and language technologies. Their perspectives are crucial for ensuring cultural and linguistic diversity in AI development. Third, workers and labour groups, especially in sectors affected by automation and algorithmic management, are underrepresented. Their input is essential to understand the socio-economic impacts of AI on employment conditions, job displacement, and workplace surveillance. Fourth, youth voices and educators are often not systematically included, even though younger generations are among the most affected by long-term AI developments and will shape its future use. To improve participation, several measures can be implemented: * Establish funded participation programs to support representatives from low-resource settings. * Create regional consultation hubs to decentralize engagement and reduce travel and access barriers. * Introduce digital participation platforms with multilingual support to broaden accessibility. * Ensure quota-based or balanced representation mechanisms in advisory groups and forums. * Strengthen partnerships with local civil society organizations that can act as intermediaries for marginalized groups. Additionally, embedding participation requirements into the structure of the Dialogue—rather than treating inclusion as optional—would help ensure sustained and meaningful engagement. Overall, improving representation is essential for ensuring that global AI governance reflects diverse social, economic, and cultural realities, rather than a narrow subset of global perspectives.

Innovative formats can significantly enhance the effectiveness of the Global Dialogue on AI under the United Nations by making interactions more inclusive, evidence-based, and action-oriented. One promising format is hybrid "policy-lab" sessions, where policymakers, researchers, and industry experts jointly work on concrete governance problems (e.g., AI safety standards or risk classification). Unlike traditional debates, these sessions would focus on producing actionable outputs such as draft guidelines or toolkits. Another effective approach is the use of continuous digital deliberation platforms. These would allow stakeholders to contribute asynchronously through structured discussions, voting mechanisms, and collaborative document editing, enabling broader participation beyond in-person meetings. The Dialogue could also benefit from scenario-based foresight workshops, where participants explore future AI developments (e.g., autonomous systems, synthetic media ecosystems) and jointly assess risks and policy responses. This helps shift governance from reactive to anticipatory approaches. A further innovation would be multi-stakeholder "regulatory sandboxes" at the international level, where countries and companies test AI governance approaches in controlled environments, sharing results with the global community. In addition, interactive science-policy interfaces, supported by the independent international scientific panel on AI, could translate complex technical findings into accessible policy insights through visual tools, simulations, and live briefings. Finally, rotating regional dialogue hubs could ensure geographical diversity and allow different regions to host thematic sessions focused on their specific priorities and challenges. Together, these formats would transform the Dialogue from a traditional intergovernmental forum into a dynamic ecosystem of continuous learning and co-creation, improving responsiveness to the fast-moving nature of artificial intelligence while ensuring inclusive global participation.

Several existing regulatory measures, governance practices, and multi-stakeholder platforms already provide useful models for strengthening global AI governance under the framework of the United Nations. One of the most advanced examples is the risk-based regulatory approach of the European Union AI Act, which classifies AI systems according to levels of risk and imposes proportionate obligations. This model is increasingly influential globally because it links legal requirements to concrete risk categories such as high-risk applications, transparency obligations, and prohibited uses. Another important framework is the OECD AI Principles, which promote values such as transparency, robustness, accountability, and human-centred design. These principles have been widely adopted as a soft-law standard and serve as a bridge between regulatory systems. In terms of practical governance mechanisms, AI safety evaluations and model auditing practices developed by leading research institutions and companies provide concrete tools for assessing system behaviour before deployment. These include red-teaming exercises, bias testing, and external audits. Multi-stakeholder platforms such as the Global Partnership on Artificial Intelligence (GPAI) demonstrate how governments, industry, and academia can collaborate on applied AI governance challenges, including responsible innovation and data governance. In addition, open-source AI ecosystems and model transparency initiatives help improve accountability by enabling independent scrutiny of algorithms and training data. However, they must be balanced with safeguards to prevent misuse. Finally, regulatory sandboxes, used in several jurisdictions, allow innovators to test AI systems under regulatory supervision, enabling experimentation while maintaining oversight and safety. Together, these approaches illustrate a combined toolkit: binding regulation, soft-law principles, technical standards, and collaborative platforms. When integrated effectively, they can form a more coherent and adaptive global AI governance system.