Bureau of Human Rights and Justice

In my opinion, the first Global Dialogue on AI Governance will be a success if it delivers practical and inclusive outcomes, not only general discussion. First, it should create a shared understanding that AI governance must support innovation, while also protecting human rights, dignity, safety and social justice. Second, it should help build common principles that can be used across countries, especially on transparency, accountability, human oversight, risk management and non-discrimination. Also, a successful dialogue should give strong attention to the needs of developing countries, civil society actors and underrepresented communities. Many countries still do not have the same technical capacity, regulatory readiness or data infrastructure. For this reason, the Dialogue should promote international cooperation, knowledge sharing and capacity building, so that AI governance does not increase global inequality. Another important outcome is to encourage interoperability between governance approaches. Different legal and policy systems will continue to exist, but they should still be able to work together around common safeguards and shared public interest goals. .. And finally, the dialogue should produce a concrete roadmap for continued action. This may include follow-up mechanisms, multistakeholder working groups, best-practice exchange, and measurable commitments. I actually wanna say that success means moving from principles to implementation, from fragmented efforts to coordinated action, and from technology-centered debate to people-centered governance that serves humanity.

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

From my perspective, these four areas are the most urgent priorities for active engagement. Protection and promotion of human rights is essential because AI systems can directly affect dignity, equality, privacy, freedom of expression, access to services and protection from discrimination. As a representative of a human rights and justice organization, I believe governance must ensure that AI serves people and does not weaken fundamental rights. Transparency, accountability and human oversight are also very very critical. In many sectors, AI-driven decisions influence employment, finance, health, education, migration and public services. People must know when AI is used, understand its impact, and have access to human review and remedy when harm occurs. AI capacity-building is another urgent priority, especially for developing countries, civil society and public institutions. Effective governance is not possible without technical knowledge, institutional readiness, data literacy and access to expertise. Capacity-building helps make AI governance more inclusive and globally balanced. Safe, secure and trustworthy AI is equally important because AI systems must be reliable, robust and resilient against misuse, harmful bias, manipulation and security risks. Trust cannot be built only through innovation; it also requires safeguards, testing, monitoring and responsible deployment. Together, these priorities support a human-centered, equitable and actionable approach to AI governance.

Yes, I believe there are several cross-cutting and emerging issues that deserve stronger attention. One important issue is global inequality in compute, data and infrastructure access. AI development is becoming concentrated in a limited number of companies and countries. Without international action, this may deepen existing digital divides and reduce the ability of many nations to participate meaningfully in AI innovation and governance. A second issue is the environmental impact of AI, including energy consumption, water use and material extraction linked to large-scale computing infrastructure. Sustainable AI governance should consider not only social and economic effects, but also environmental responsibility. A third issue is access to remedy and redress for AI-related harms. Governance discussions often focus on principles, but affected individuals and communities also need practical complaint mechanisms, independent oversight and pathways to justice when AI causes discrimination or harm. Another emerging issue is the effect of AI on information integrity and democratic trust, including deepfakes, synthetic media, automated manipulation and large-scale disinformation. These risks can damage public trust, social cohesion and democratic participation. Finally, there should be more discussion on meaningful participation from civil society, grassroots communities and the Global South in shaping AI rules. Inclusion should not be symbolic; it should be built into decision-making structures. AI governance will be more legitimate and effective when those most affected are also meaningfully represented.

The governance gaps and recent developments in the thematic areas I selected are affecting my sector in a very direct way, especially in data science, business intelligence, predictive analytics and human rights advocacy. AI adoption is moving faster than governance capacity. Many organizations are already using AI for decision support, automation, profiling and content generation, but clear standards on accountability, explainability, data quality and human oversight are still uneven across countries and sectors. One major challenge is the risk of biased or low-quality data leading to unfair outcomes, especially for vulnerable populations. In both public-interest and nonprofit contexts, this can affect access to services, inclusion, and trust. Another challenge is the lack of harmonized governance approaches across regions. Different legal, technical and ethical frameworks make cross-border collaboration more difficult, especially for International Organisations working on human rights, justice and development. At the same time, there are important opportunities. Recent advances in AI can improve evidence-based decision-making, early warning systems, multilingual access to information, and more efficient public-service delivery. In the nonprofit and human rights space, AI can support research, case analysis, trend detection and better targeting of limited resources. In business intelligence, it can strengthen forecasting, operational efficiency and strategic planning. However, these opportunities can only be sustainable if governance keeps pace. Capacity-building is essential, especially for civil society and institutions in lower-resource settings. There is also a strong need for transparent risk assessment, meaningful human oversight, and practical mechanisms for remedy when harm occurs. In my view, the greatest opportunity is to build inclusive AI governance that allows innovation to serve society, while the greatest risk is to let governance remain fragmented, reactive and unequal.

In my opinion, the AI Dialogue can play a very important role as a trusted multilateral space where governments, civil society, academia, the private sector and technical experts can exchange views on AI governance in an inclusive manner. The main value of the Dialogue is not to replace national or regional frameworks, but to help create common understanding, shared language and practical cooperation across different systems. The UN describes the Dialogue as an inclusive platform to discuss critical AI issues and to foster interoperability between national approaches under Resolution 79/325... The Dialogue can advance international cooperation by helping identify minimum common principles on safety, human rights, transparency, accountability and human oversight. It can also support knowledge sharing on what is working in practice, including risk management, public-sector deployment, impact assessment and capacity building. This is especially important because many countries and institutions do not have equal technical, legal or financial capacity to govern AI effectively Another important role is to ensure that voices from developing countries, civil society and underrepresented communities are meaningfully included. International cooperation on AI governance will not be credible if it is shaped only by a small number of powerful actors. The Dialogue can help reduce fragmentation by connecting existing efforts, encouraging policy interoperability and promoting more balanced participation across regions I think, in this way, the AI Dialogue can become a bridge between principle and implementation, and between innovation and protection of the public interest.

The AI Dialogue should build on existing international initiatives instead of starting from zero. Important foundations already exist in the UN system and beyond. These include the Global Digital Compact, which set the broader framework for digital cooperation and AI governance, and the UN High-Level Advisory Body on AI, which already developed globally inclusive thinking on governance options. It should also connect with UNESCO's Recommendation on the Ethics of AI and its readiness assessment work, which provide practical experience on ethics, governance and institutional capacity across many countries. Beyond the UN, the Dialogue should connect with the OECD AI Principles, the G7 Hiroshima AI Process, GPAI, and the Internet Governance Forum, especially because the UN has indicated that the Dialogue complements existing forums such as the IGF rather than duplicating them.These mechanisms already contain valuable lessons, technical tools and policy experience. The added value of the AI Dialogue is its universal legitimacy and inclusiveness. Unlike smaller or regionally limited initiatives, the Dialogue can bring all countries and stakeholder groups into one space, including actors from the Global South, civil society and human rights communities who are not always equally represented elsewhere. Its unique contribution can be to connect fragmented efforts, identify shared priorities, support interoperability across governance models, and translate broad principles into more practical cooperation, capacity-building and follow-up action.

Different stakeholders should contribute according to their real experience and responsibilities. Governments can share policy lessons, regulatory challenges and national priorities. Civil society can bring the perspective of affected communities, rights protection and public accountability. Academia and technical experts can provide evidence, research and independent analysis. The private sector can contribute practical knowledge on development, deployment, risk management and innovation. International Organisations can help connect efforts and support coordination. For the format, the Dialogue should be inclusive, practical and action-oriented. It should combine high-level plenary sessions with smaller thematic working groups where participants can discuss concrete issues in depth. There should also be space for written inputs, regional consultations and hybrid participation, so that stakeholders who cannot travel can still engage meaningfully. The structure should ensure balanced representation across regions and sectors, with clear moderation and transparent documentation of outcomes. It would also be useful to include short policy labs or case-based sessions focused on real governance challenges. In my view, the Dialogue should not be only a speaking forum. It should produce clear follow-up steps, summary recommendations and mechanisms for continued multistakeholder cooperation.

Several important voices remain underrepresented in global AI governance discussions. These include stakeholders from the Global South, small and lower-resource states, grassroots civil society organizations, youth, persons with disabilities, indigenous communities, linguistic minorities, refugees, and communities already affected by inequality, surveillance or digital exclusion. Too often, global AI debates are dominated by governments, large technology companies and institutions with more financial and technical resources. This creates a risk that governance frameworks will not fully reflect the realities of those most affected by AI harms or those with the least capacity to respond. These groups can be included through concrete measures, not only symbolic invitations. This means financial support for participation, multilingual access, hybrid engagement, accessible formats, early consultation before decisions are drafted, and dedicated speaking opportunities in formal sessions. It is also important to include community-based organizations in working groups and follow-up processes. In my opinion, AI governance will be more legitimate, fair and effective when participation is broadened from expert-centered discussion to truly inclusive representation.

To foster meaningful and dynamic engagement, the AI Dialogue should use formats that go beyond traditional speeches. Interactive roundtables, policy labs and scenario-based workshops can help participants discuss concrete governance challenges and compare practical solutions. These formats are useful because they encourage exchange, problem-solving and joint learning. Another effective format is multistakeholder case discussion, where governments, civil society, technical experts and private-sector actors respond together to real or realistic AI governance situations, such as bias in public services, cross-border data use, deepfakes or accountability failures. This can make the Dialogue more practical and less abstract. Regional breakout sessions could also be valuable, allowing participants to highlight different local realities while still feeding into a shared global discussion. In addition, digital participation tools such as moderated online consultations, live polling, collaborative drafting spaces and structured written submissions can help broaden participation before, during and after the event. In my view, the most effective format is one that combines inclusiveness with practical output. The Dialogue should create spaces where participants not only speak, but also co-develop recommendations, identify shared priorities and build lasting cooperation.

Examples already exist at policy, institutional and technical levels. The EU AI Act is an important example because it introduces a risk-based approach, with stronger obligations for high-risk AI systems. UNESCO's Recommendation on the Ethics of Artificial Intelligence offers a global normative framework grounded in human rights, inclusion and accountability. The OECD AI Principles also provide useful guidance on trustworthy AI, transparency and responsible innovation. At the practice level, algorithmic impact assessments, independent audits, documentation standards, and human oversight mechanisms are concrete tools that can reduce harm and improve accountability. In the technical and collaborative space, open standards, model documentation, data governance frameworks, and secure testing environments can support safer and more interoperable AI systems. In my view, the most effective approaches are those combining regulation, ethics, technical safeguards and capacity-building. Good AI governance should not rely on one instrument only. It should connect law, institutional responsibility, and practical implementation in a way that protects people while still supporting innovation.