Ministry of Justice and Human Rights

The success of the first Global Dialogue on AI Governance will depend on moving beyond general statements to deliver practical, inclusive, and measurable results. First, it should establish a shared foundation of principles for trustworthy AI, including human rights, transparency, accountability, safety, fairness, privacy, and meaningful human oversight. While regulatory models may differ, aligning on core values can reduce fragmentation and strengthen international trust. Second, the Dialogue must amplify the voices of developing countries, particularly from Africa and other underrepresented regions. AI governance should not be dominated by a few powerful nations or tech corporations. Ensuring meaningful participation from governments, civil society, academia, youth, and the private sector is essential to creating rules that reflect diverse realities. Third, concrete cooperation mechanisms should be created, such as commitments to capacity building, technical assistance, knowledge sharing, and support for digital infrastructure. Many nations want to govern AI responsibly but lack expertise, computational resources, data systems, or regulatory readiness. Bridging these gaps is critical. Fourth, the Dialogue should address urgent cross border risks, including misinformation, algorithmic discrimination, cyber misuse, and market concentration. Even without full consensus, identifying priority risks and launching joint working tracks would represent meaningful progress. Fifth, success means recognizing AI not only as a risk to manage, but also as a tool for development. The Dialogue should highlight how AI can advance education, healthcare, agriculture, justice, and public services, especially in the Global South. Finally, the most important outcome would be continuity, a roadmap with clear milestones, multi stakeholder working groups, and a commitment to regular review. A successful first Dialogue should launch an ongoing governance process, not conclude with a single event. In short, success should be measured by inclusion, actionable commitments, developmental impact, and sustained global cooperation.

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

I selected these priorities because effective AI governance must be both principled and practical, especially for developing countries seeking to harness AI for inclusive growth and public service transformation. AI capacity building is essential because many countries still face gaps in technical expertise, institutional readiness, digital infrastructure, and regulatory capacity. Without investment in skills, research ecosystems, and public sector capabilities, global participation in AI governance will remain unequal. Protection and promotion of human rights is a fundamental priority. AI systems can affect privacy, equality, freedom of expression, access to justice, and non discrimination. Governance frameworks must ensure that innovation advances human dignity and does not reinforce exclusion or bias. Transparency, accountability, and human oversight are critical for public trust. Citizens should know when AI is used, understand the basis of important decisions, and have access to remedies when harm occurs. Human oversight remains necessary, particularly in high impact sectors such as healthcare, education, employment, and public administration. Safe, secure and trustworthy AI is equally urgent in a context of rapidly expanding deployment. AI systems should be reliable, resilient, and protected against misuse, cyber threats, misinformation, and harmful unintended consequences. Safety measures should be integrated throughout the lifecycle of AI systems. Together, these priorities create a balanced approach, building national capacity, protecting rights, strengthening accountability, and managing risks. They are particularly relevant for countries in the Global South, where AI presents major opportunities to improve education, agriculture, healthcare, justice, and digital public services, but where governance gaps can also deepen inequalities if left unaddressed.

Yes. While the listed themes cover many core priorities, several important cross cutting and emerging issues deserve stronger attention. A key issue is digital inequality and compute access. AI development increasingly depends on access to computing power, cloud infrastructure, quality datasets, and advanced chips. Many developing countries risk exclusion from the AI economy due to limited infrastructure and high costs. Governance discussions should address equitable access to compute resources and digital infrastructure. Another priority is data governance and data sovereignty. Beyond open data, there is a need for fair rules on data ownership, cross border data flows, consent, localization, and community data rights. Countries need frameworks that enable innovation while protecting citizens and national interests. Equally important is environmental sustainability. Large AI systems can consume significant energy and water resources. AI governance should include environmental accountability, energy efficiency, green computing standards, and sustainable infrastructure planning. We must also consider the labour market transition and the future of work. AI will reshape jobs, skills demand, and productivity. More attention is needed on workforce reskilling, youth employment, social protection, and supporting workers affected by automation. Linguistic and cultural inclusion for underrepresented regions is another critical area. Many languages, especially African and indigenous languages, remain poorly represented in AI systems. Governance should encourage multilingual datasets, inclusive language technologies, and cultural diversity in AI design. The concentration of market power also demands focus. A small number of actors control key AI models, compute resources, and platforms. This can limit competition and innovation. Fair market access and open innovation ecosystems should be part of governance discussions. Finally, AI in public sector decision making requires special safeguards. As governments adopt AI for administration, justice, security, and welfare delivery, we need stronger measures for due process, procurement standards, auditability, and citizen redress mechanisms. These issues cut across development, rights, sustainability, and global equity, making them essential to future AI governance.

In my country, Togo, and across many parts of Africa, governance gaps in the areas of capacity building, human rights, accountability, and trustworthy AI present both significant challenges and meaningful opportunities. A major challenge is limited institutional and technical capacity. Many public institutions are interested in using AI to improve services, but often lack specialized expertise, regulatory frameworks, data governance systems, and adequate digital infrastructure. This can slow innovation or lead to poorly governed adoption. Another concern is the risk to human rights and public trust. Without clear safeguards, AI systems used in recruitment, security, finance, or public administration may reinforce bias, exclude vulnerable populations, or undermine privacy. In contexts where digital literacy is uneven, opaque systems can reduce citizens' trust in institutions. Transparency and accountability gaps are also significant. Many imported AI tools function as "black boxes," making it difficult for governments to assess how decisions are made, who is responsible for errors, and whether systems comply with national laws or public values. Public procurement processes often lack AI specific standards for auditing, explainability, and oversight. At the same time, the opportunities are substantial. AI can strengthen education, agriculture, healthcare, justice, language inclusion, and administrative efficiency. In Africa, AI powered tools can help bridge service delivery gaps, improve early warning systems, optimize resource allocation, and support underserved communities. There is also a strategic opportunity to build locally relevant and inclusive AI ecosystems. African countries can develop solutions in local languages, create context aware governance models, and encourage startups, universities, and public institutions to collaborate. For the public sector, trustworthy AI can accelerate digital transformation if deployed responsibly. The most significant need is to convert today's governance gaps into readiness through skills development, regulatory innovation, regional cooperation, and citizen centered safeguards. If managed well, AI can become a driver of inclusive development rather than a source of new inequality.

The AI Dialogue can play a critical role as a trusted multilateral platform for building convergence, inclusion, and practical cooperation on AI governance. Because AI impacts all countries but capacities and interests differ, no single state or company can govern it alone. The Dialogue can help bridge this gap. It can promote a shared understanding of risks, opportunities, and baseline principles such as safety, human rights, transparency, accountability, and human oversight. While national approaches may vary, greater alignment can reduce fragmentation and regulatory uncertainty. The Dialogue can also strengthen inclusive participation, ensuring that developing countries, small states, civil society, academia, youth, and the private sector have a meaningful voice. International AI governance will be more legitimate and effective when it reflects diverse realities, not only the perspectives of major powers or large technology firms. Another key function is to catalyze capacity building partnerships. Many countries need support in policy design, regulatory readiness, digital infrastructure, data governance, and technical skills. The Dialogue can connect countries with expertise, financing, training networks, and South South as well as North South cooperation. It can further encourage interoperability between governance approaches. Different legal and policy models are emerging worldwide. The Dialogue can help identify common standards, compatible practices, and areas for mutual recognition while respecting national sovereignty. The platform can also coordinate action on cross border challenges such as misinformation, cyber misuse, market concentration, deepfakes, and risks from advanced AI systems. These issues require joint responses, information sharing, and early warning cooperation. Finally, the Dialogue can serve as an implementation mechanism, not only a discussion space, by launching working groups, voluntary commitments, knowledge platforms, and measurable follow up processes. In short, the AI Dialogue can transform fragmented global debates into sustained cooperation that is practical, equitable, and development oriented, helping ensure that AI benefits all countries and communities.

The AI Dialogue should build upon and connect with existing international, regional, and multi stakeholder initiatives rather than duplicate them. A strong starting point is UNESCO's Recommendation on the Ethics of Artificial Intelligence, which provides globally endorsed principles on human rights, fairness, transparency, and accountability. The Dialogue can help translate these principles into implementation support and peer learning. It should also connect with the UN Global Digital Compact process and broader UN digital cooperation efforts, which emphasize inclusive and rights based digital governance. Alignment with these frameworks would strengthen coherence across the multilateral system. The OECD AI Principles and the Global Partnership on AI offer practical policy guidance, research networks, and evidence based approaches. The Dialogue could broaden access to these resources for countries that are not fully represented in existing forums. Regional initiatives are equally important, including the African Union Continental AI Strategy, European Union AI Act, and emerging national AI strategies across Asia, Latin America, and the Middle East. These experiences provide valuable lessons on regulation, innovation, and institutional readiness. The Dialogue should also engage technical and governance communities such as the Internet Governance Forum, standards bodies, open source communities, academia, and industry alliances. AI governance requires technical expertise alongside policy leadership. Its added value would be fourfold. One key contribution is universal legitimacy as a UN linked space where all countries, especially developing nations, can participate equally. Another is its ability to bridge fragmentation by connecting separate regional and sectoral efforts into a more coherent global conversation. The Dialogue can also advance capacity building and implementation, helping countries move from principles to practice through training, toolkits, and partnerships. Finally, it can maintain a strong development focus, ensuring AI governance supports education, health, agriculture, public services, and inclusion, especially in the Global South. In short, the AI Dialogue can become the connector, amplifier, and implementation bridge across today's fragmented AI governance ecosystem.

Different stakeholders should contribute according to their comparative strengths, within an open and balanced multistakeholder framework. Governments can share policy experiences, national strategies, regulatory lessons, and public sector use cases. They can also identify capacity needs and opportunities for international cooperation. Private sector actors can contribute technical expertise, safety practices, transparency measures, responsible innovation models, and investment partnerships. Because many frontier AI systems are privately developed, their engagement is essential. Academia and research institutions can provide independent evidence, risk assessments, benchmarking methodologies, and foresight on emerging trends. Civil society and human rights organizations can ensure that inclusion, consumer protection, labor rights, gender equality, and vulnerable communities remain central to the discussion. Technical communities and open source ecosystems can advise on standards, interoperability, cybersecurity, open models, and practical implementation pathways. Youth and underrepresented regions, especially from Africa and the Global South, should be meaningfully included, not symbolically represented, since AI governance decisions will shape their future development. - Recommendations for format and structure A hybrid and accessible format is recommended, combining in person and virtual participation, multilingual interpretation, remote interventions, and publicly available documents to widen access. A multi track structure could organize sessions around key themes such as safety, human rights, capacity building, innovation, and development impacts. Action oriented working groups should be created, forming smaller intersessional groups that continue work between annual meetings and produce practical outputs. Regional consultations are important, holding preparatory dialogues through regional bodies such as the African Union, European Union, and others to reflect local priorities. Evidence and case study driven sessions are advised, using real implementation examples from healthcare, education, agriculture, justice, and public administration. Each Dialogue should produce clear deliverables, ending with a summary of recommendations, voluntary commitments, partnerships, and next step milestones. To maintain the balance and legitimacy of the Dialogue, safeguards must be in place to ensure that no single bloc whether states, companies, or regions can dominate the agenda setting.

Several important voices remain underrepresented in global discussions on AI governance, which can weaken legitimacy, fairness, and practical relevance. A significant gap exists for developing countries, particularly in Africa, small island states, and least developed nations, which are often underrepresented despite being profoundly impacted by AI driven economic and social shifts. Many face barriers related to funding, technical capacity, and limited access to global forums. Furthermore, local communities and end users who interact with AI systems in their daily lives are rarely heard. Individuals such as farmers, teachers, health workers, informal laborers, persons with disabilities, and citizens accessing public services can offer invaluable, practical insights into how AI affects livelihoods and rights. Youth perspectives, while sometimes invited, are often included symbolically rather than meaningfully, despite the fact that younger generations will live with the long term consequences of today's AI governance decisions. Linguistic and cultural minorities, including speakers of African and indigenous languages, are frequently excluded from conversations that are predominantly conducted in English and shaped within a narrow set of digital cultural contexts. Additionally, small and medium sized enterprises, startups, and researchers from the Global South remain less visible compared to large multinational technology firms, even though they play a critical role in fostering locally relevant innovation. Finally, labor organizations and workers' representatives require stronger inclusion as AI transformation reshapes employment, productivity, and workplace rights globally. How they could be included : - Establishing dedicated participation funding for travel, connectivity, and fellowships is essential to lower barriers to entry. - Adopting hybrid and multilingual formats with professional interpretation and accessible digital platforms can dramatically widen engagement. - Leveraging regional consultation mechanisms through bodies like the African Union, ASEAN, and other regional forums can help ground discussions in local realities. - Creating reserved speaking and leadership roles for underrepresented groups within panels, drafting committees, and steering bodies ensures their voices shape outcomes. - Implementing open calls for written submissions and community case studies helps capture grassroots experiences and practical knowledge. - Developing youth co creation models, where young experts are involved in agenda setting and decision making processes, moves beyond tokenistic participation. - Providing support for local research ecosystems and encouraging South South knowledge exchange strengthens capacity and fosters home grown expertise. Inclusive AI governance requires moving from passive representation to active influence. Broader and more meaningful participation will ultimately lead to smarter, fairer, and more globally legitimate outcomes.

Meaningful engagement during the AI Dialogue will require formats that move beyond traditional speeches and create interactive, solution-oriented participation. Several innovative approaches could be especially effective. - Policy Labs and Co-Creation Workshops : Small multi-stakeholder groups can work on real governance challenges such as AI procurement rules, deepfake response frameworks, or human rights safeguards. These sessions should produce draft recommendations or toolkits. - Scenario Simulations and Tabletop Exercises : Participants could simulate responses to cross-border AI incidents, such as misinformation during elections, AI-enabled cyberattacks, or failures in public-sector AI systems. This helps decision-makers test cooperation mechanisms in practice. - Regional Voices Roundtables : Dedicated sessions for Africa, Latin America, small island states, and least developed countries can surface priorities often missed in global debates. Outcomes should feed directly into plenary discussions. - Solutions Marketplace : Governments, startups, universities, and civil society organizations could showcase practical AI tools for health, education, agriculture, language inclusion, and public administration. This balances risk discussions with development opportunities. - Youth and Future Generations Forum : Young professionals, students, and innovators should have structured agenda-setting roles, not only speaking slots. Their recommendations could be formally presented to the main Dialogue. - Open Microphone / Lightning Interventions : Short timed interventions from diverse participants can widen participation beyond invited panelists and reduce gatekeeping. - Evidence Clinics : Researchers and technical experts can present concise evidence briefs on topics such as frontier model risks, labor impacts, or AI energy use, helping policymakers make informed decisions. - Digital Participation Platforms : Use real-time multilingual polling, collaborative drafting tools, remote breakout rooms, and public consultation portals to include those unable to attend physically. - Commitment Sessions : Stakeholders announce voluntary commitments, partnerships, training initiatives, or governance pilots, followed by progress tracking before the next Dialogue. The most effective format would combine plenaries for vision-setting, workshops for problem-solving, and continuous digital participation for inclusion and accountability.

Several existing policies, practices, and approaches offer useful models for effective AI governance and practical risk management. - Risk based regulatory approaches : The European Union AI Act is a leading example. It classifies AI systems by levels of risk and applies stronger obligations to high risk uses such as critical infrastructure, employment, and public services. This helps focus regulation where harm is greatest. - Ethical governance frameworks : UNESCO Recommendation on the Ethics of Artificial Intelligence provides global guidance on human rights, fairness, transparency, environmental sustainability, and inclusion. It is valuable because it applies to countries at different levels of development. - Algorithmic impact assessments : Before deploying AI in the public sector, governments can require impact assessments that evaluate bias, privacy, accountability, and social consequences. This proactive model is especially useful for welfare, policing, justice, and education systems. - Human in the loop decision systems : For high impact decisions, AI should support not replace human judgment. Human review, appeal mechanisms, and clear responsibility lines improve fairness and public trust. - AI procurement standards : Governments increasingly need procurement rules requiring explainability, cybersecurity, auditability, data protection, and vendor accountability when purchasing AI systems. - Regulatory sandboxes : Controlled testing environments allow innovators and regulators to experiment safely with new AI tools while learning what rules are effective. This can help startups and emerging markets innovate responsibly. - Open source and local innovation ecosystems : Open source AI tools, open datasets, and public interest digital infrastructure can lower barriers to entry and help developing countries build locally relevant solutions, including tools in African languages. - Capacity building partnerships : Collaboration between governments, universities, civil society, and industry can train regulators, judges, civil servants, and developers on AI governance. An effective approach combines rules, technical safeguards, public accountability, and innovation support. No single model is sufficient; successful governance requires adapting these tools to national contexts while maintaining shared global principles.