Ministry of Information Technology, Communication and Innovation - AI UNIT

For Mauritius, a Small Island Developing State navigating AI's transformative potential, the first Global Dialogue's success hinges on four interconnected outcomes: 1. Inclusive Governance Framework Success requires that AI governance reflects the priorities of all nations, not merely the technologically advanced. The Dialogue must establish principles: fairness, accountability, inclusiveness and responsibility that are universally applicable yet context-sensitive. Mauritius has embedded these principles (the FAIR framework) into its National AI Strategy precisely because governance must be adaptable to different economies, capacities and development stages. 2. Bridging the Capacity Divide The Dialogue must translate commitments into concrete mechanisms for capacity-building, particularly for developing nations. This includes access to computational infrastructure, skills development, open-source AI models and data governance support. Without deliberate action to bridge these divides, AI will amplify existing inequalities rather than unlock shared prosperity. 3. Human-Centred Safeguards Success means establishing that AI systems, regardless of where they are developed, must remain subject to meaningful human oversight, transparency and accountability. High-impact applications affecting rights, services and democratic decision-making require robust safeguards that protect citizens while enabling responsible innovation. 4. Shared Regional and International Cooperation The Dialogue should catalyse practical cooperation mechanisms: peer learning networks, technical assistance partnerships and interoperable governance approaches. For small states like Mauritius, regional coordination (through SADC, AU, COMESA frameworks) is essential to avoid fragmentation and build collective resilience. Ultimately, success is measured by whether the Dialogue creates the conditions for every nation to govern AI responsibly within its own context, while contributing to global standards that protect human dignity and advance the Sustainable Development Goals.

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

Safe, secure and trustworthy AI Mauritius has launched comprehensive FAIR Guidelines (Fairness, Accountability, Inclusiveness and Responsibility) that establish risk-based governance from system design through decommissioning. We are committed to ensuring that AI adoption strengthens institutional trust rather than undermining it, with particular focus on high-impact applications in public services, healthcare and financial systems. AI capacity-building As a Small Island Developing State, capacity-building is foundational to our National AI Strategy. We are investing in institutional literacy across government, skills development for our workforce and support for SMEs and technopreneurs. Without deliberate capacity investment, developing nations cannot meaningfully participate in or benefit from AI governance. This is urgent for us and critical for global equity. Transparency, accountability and human oversight Our FAIR framework explicitly requires clear accountability structures, meaningful human oversight (particularly in high-risk decisions) and proportionate transparency throughout the AI lifecycle. We believe that AI systems should support and not replacec human judgement and that responsibility must remain visible and enforceable at all stages. Protection and promotion of human rights Our National Strategy affirms that AI must strengthen democratic governance, protect constitutional rights and ensure no citizen is left behind. We prioritize safeguards against discrimination, respect for dignity and meaningful participation in decisions affecting citizens' rights and access to essential services. These four priorities are mutually reinforcing and central to Mauritius's vision of responsible, human-centred AI that serves the public good while bridging development divides. They directly advance the Global Dialogue's core mission: ensuring all nations shape AI governance and share its benefits equitably.

Yes. While General Assembly Resolution 79/325 captures essential thematic areas, Mauritius identifies three cross-cutting issues requiring urgent dialogue at the Global Dialogue: 1. AI Governance for Small Island Developing States (SIDS) and Ocean States The listed themes address AI challenges generically, but SIDS face distinctive vulnerabilities: geographic insularity, limited computational infrastructure, brain drain, climate vulnerability and economic dependence on specific sectors (tourism, fisheries, financial services). AI governance frameworks must account for these realities. For ocean states like Mauritius, AI applications in maritime surveillance, fisheries management, climate modelling and disaster preparedness are existential priorities, yet little international guidance addresses these sector-specific governance needs. The Global Dialogue must ensure that SIDS-specific contexts shape universal frameworks, not merely adapt them after the fact. 2. Data Sovereignty and Equitable Transnational Data Flows The theme "open data and open AI models" is important, but it does not adequately address the tension between data openness and national sovereignty. Developing nations face pressure to share data for AI development while losing control over their own information assets and the economic value they generate. The Global Dialogue must establish equitable data governance frameworks that protect developing nations' interests while enabling beneficial knowledge-sharing. This is essential to ensuring benefits are "shared by all," not concentrated in data-rich nations. 3. AI's Role in Democratic Governance and Institutional Integrity As AI systems increasingly influence public administration, electoral processes and governance decisions, the Dialogue must address how AI strengthens rather than undermines democratic accountability. This includes safeguards against automated decision-making that obscures political choice, ensures public participation in governance and protects against manipulation or surveillance that erodes democratic trust. For all nations to "govern this transformation together," democratic safeguards must be explicit in the Dialogue's framework. These cross-cutting issues affect how other themes are implemented. Without addressing SIDS contexts, data equity and democratic safeguards, global AI governance risks replicating existing power imbalances and leaving developing nations as consumers rather than shapers of AI futures. The Global Dialogue's mission demands their inclusion.

Governance Gaps & Regional Impact: Mauritius faces critical governance gaps as rapid AI adoption outpaces regulatory capacity. Key challenges include: Regulatory fragmentation: No unified AI governance framework across sectors (finance, healthcare, public services) Capacity constraints: Limited technical expertise in AI governance among regulators and policymakers International alignment: Balancing sovereign governance with participation in global standards without imposing unsuitable models SIDS vulnerabilities: Limited resources for monitoring and enforcement; susceptibility to imported AI risks Opportunities: Leapfrogging potential: Adopting responsible AI governance early, positioning Mauritius as a regional leader in the Indian Ocean and Africa FAIR Guidelines implementation: Using your non-binding guidelines to create adaptive, sector-responsive governance without premature legislation Regional collaboration: Strengthening partnerships with regional peers facing similar challenges Inclusive governance model: Leveraging your shared governance approach across public, private, academic, and civil society sectors Investment attraction: Demonstrating robust AI governance to attract responsible AI investment and innovation Most Significant Priority: Operationalizing the FAIR Guidelines as a phased governance foundation—enabling innovation while establishing safeguards progressively as capacity grows.

The Global Dialogue on Artificial Intelligence Governance can serve as a critical platform for: 1. Bridging the Governance Gap The Dialogue can facilitate knowledge-sharing between advanced AI nations and developing countries, ensuring AI governance frameworks are contextually appropriate rather than one-size-fits-all. For Small Island Developing States like Mauritius, this creates space to articulate unique vulnerabilities—limited regulatory capacity, economic constraints and exposure to imported AI risks—while learning from global best practices. 2. Operationalizing Inclusive Governance Models The Dialogue can advance shared governance approaches that recognize the complementary roles of governments, private sector, academia and civil society. This enables countries to move beyond traditional regulatory models toward adaptive, multi-stakeholder frameworks that reflect diverse national capacities and priorities. 3. Establishing Compatible Standards Without Hegemony Rather than imposing global rules, the Dialogue can develop principles and best practices that allow nations to maintain sovereignty while ensuring interoperability. This is essential for SIDS and developing nations that cannot absorb duplicative or contradictory regulatory requirements. 4. Supporting Capacity Building and Technical Cooperation The Dialogue can catalyze concrete partnerships—bilateral technical assistance, international training programs, access to expertise and researcher exchanges—enabling developing countries to build AI governance capacity progressively. 5. Amplifying Underrepresented Voices The Dialogue provides a platform for Small Island Developing States and other vulnerable nations to articulate their interests in global AI governance, ensuring international cooperation reflects diverse developmental contexts and priorities. 6. Creating Accountability Mechanisms A structured dialogue can establish peer-review processes, progress monitoring and shared learning platforms that hold all nations accountable to commitments on responsible AI governance.

1. UNESCO Recommendation on AI Ethics (2021) Provides foundational ethical principles (transparency, accountability, human rights). The Dialogue can operationalize these principles into actionable governance mechanisms and sectoral implementations. 2. OECD AI Principles & Recommendations Established governance best practices for advanced economies. The Dialogue adds value by contextualizing these for developing nations and SIDS, addressing capacity constraints and resource limitations. 3. ITU Focus on AI for Good The back-to-back co-location of the Global Dialogue with ITU's AI for Good Global Summit creates synergy—the Dialogue can translate AI's development potential into governance frameworks that enable responsible innovation in developing regions. 4. UN Sustainable Development Goals (SDGs) AI governance must align with SDG implementation. The Dialogue can establish how responsible AI governance supports development priorities while managing risks in healthcare, education, agriculture and climate action. 5. Regional AI Initiatives (African Union, ASEAN, etc.) The Dialogue can harmonize regional approaches, preventing fragmentation and enabling cross-regional learning among peers facing similar challenges. Added Value the AI Dialogue Brings: 1. Adaptive Governance Framework Development Unlike existing initiatives focused on principles or sectoral guidance, the Dialogue can develop phased, flexible governance models suited to countries at different development stages—critical for SIDS and developing nations. 2. Multi-Stakeholder Coordination The Dialogue uniquely brings governments, private sector, academia and civil society into structured dialogue, translating competing interests into shared governance approaches. 3. Bridging Implementation Gaps Existing initiatives often lack mechanisms linking principles to domestic implementation. The Dialogue can create peer-learning networks and technical cooperation pathways that accelerate governance operationalization. 4. Voice for Underrepresented Regions While OECD, UNESCO and ITU provide valuable frameworks, they center advanced economies. The Dialogue ensures developing nations and SIDS shape global AI governance norms from the outset. 5. Accountability & Progress Monitoring The Dialogue can establish periodic review mechanisms ensuring nations translate commitments into tangible governance progress, preventing governance frameworks from remaining aspirational.

Mauritius recommends a multi-layered, inclusive structure that ensures meaningful participation from all stakeholders while producing actionable outcomes and advancing the Global Dialogue's core mission of collective governance: Stakeholder Participation Architecture Governments: Should lead thematic working groups on governance implementation, with mandatory participation from regulators, not only foreign ministries. Technical expertise matters as much as diplomatic skill. Private Sector & Technopreneurs: Must contribute expertise on real-world challenges and emerging risks, but within transparent governance frameworks that prevent regulatory capture. Small and medium enterprises deserve dedicated space alongside large corporations. Academia & Research Institutions: Should provide evidence-based analysis on AI impacts, ethical implications and sector-specific governance needs. Independent research helps ground dialogue in facts rather than commercial interests. Civil Society & Community Representatives: Must represent affected populations particularly marginalised groups, SIDS and developing nations. Their voices prevent governance from becoming elite-driven and ensure the Global Dialogue reflects all nations' priorities, not just the technologically advanced. Technical Communities: Open-source developers, data scientists and technologists should guide discussions on interoperability, standards and practical implementation. Recommended Format & Structure 1. Regional Consultation Hubs Rather than Geneva-centric dialogue, establish regional hubs (Africa, Asia-Pacific, Americas, Europe) where SIDS and developing nations can articulate priorities with peer support before global deliberation. This ensures inclusive participation. 2. Sectoral Deep-Dives Beyond thematic clusters, convene sector-specific working groups (healthcare, finance, agriculture, education, public administration) where context-specific governance needs can be addressed with relevant experts. 3. Practitioner-Led Pilot Projects The Dialogue should identify and document real-world AI governance randeproduces credible, universally applicable frameworks. 4. Structured Follow-Up Mechanism Establish a permanent secretariat with dedicated support for implementation, capacity-building coordination and regular progress review. A dialogue without accountability mechanisms becomes symbolic. 5. Youth & Future-Focused Participation Include voices of those who will inherit AI governance decisions, ensuring the Dialogue shapes futures equitably across generations. Success requires genuine power-sharing, not tokenism. All stakeholders must influence outcomes, ensuring the Dialogue truly governs transformation collectively.

Mauritius identifies critical gaps in current AI governance dialogue that the Global Dialogue must address to fulfill its mandate of inclusive, equitable governance: Underrepresented Voices 1. Small Island Developing States (SIDS) and Least Developed Countries (LDCs) SIDS and LDCs are shaped by AI governance decisions yet rarely influence them. Their specific vulnerabilities:climate risk, economic dependence, limited technical capacity, are absent from mainstream discourse. These nations experience AI's risks (surveillance, data extraction, job displacement) without proportional access to its benefits. The Global Dialogue's core mission demands their voice be central, not peripheral. 2. Frontline Communities and Affected Populations AI governance is dominated by policymakers and technologists. Yet those most affected: agricultural workers facing automation, healthcare patients subject to algorithmic diagnosis, citizens in criminal justice systems using AI risk assessment have minimal voice. Their lived experience of AI's real-world impacts must inform governance ensuring benefits reach all people, not abstractions. 3. Women, Indigenous Peoples and Marginalised Groups Women remain underrepresented in AI technical and governance spaces. Indigenous communities' data rights, cultural IP and knowledge systems are rarely protected in global frameworks. Individuals, persons with disabilities and ethnic minorities whose rights are vulnerable to algorithmic bias need dedicated representation. Without their voice, the Dialogue cannot claim to represent all nations and communities. 4. Civil Society from the Global South Northern NGOs dominate civil society participation. Southern civil society organisations—who understand local contexts and power dynamics—lack resources and access to governance forums. 5. Technopreneurs and SMEs in Developing Nations Global AI dialogue centers large tech corporations. Smaller innovators from developing nations, who understand local needs and constraints, are excluded from shaping governance that affects their work. Inclusion Mechanisms 1. Funded Participation: Provide travel and capacity support for SIDS, LDC and civil society representatives. 2. Regional Hubs: Decentralise dialogue beyond Geneva/New York to enable broader participation. 3. Dedicated Consultation Spaces: Reserve sessions specifically for underrepresented communities. 4. Translated Materials & Accessibility: Ensure dialogue is linguistically and physically accessible. 5. Community-Led Research: Fund research by affected communities documenting AI's local impacts. 6. Youth & Future Generations: Institutionalise youth participation. Without deliberate inclusion, the Dialogue risks replicating power imbalances, leaving the most vulnerable without voice in governing their futures.

Mauritius proposes innovative formats that move beyond traditional conference models to generate authentic dialogue and actionable outcomes that are essential to the Global Dialogue's mission of collective governance: 1. Immersive Scenario Laboratories Rather than abstract debate, convene facilitated "AI governance labs" where diverse stakeholders simulate real governance challenges. Participants collectively design solutions, exposing tensions and building shared understanding across nations and perspectives. 2. Community-Led Case Study Showcases Commission research from SIDS, developing nations and civil society documenting how AI affects their specific contexts (fisheries management, disaster preparedness, agricultural livelihoods). Present these not as presentations but as interactive dialogues where affected communities directly engage policymakers and technologists. This ensures all nations' experiences shape the dialogue. 3. Virtual Regional Hubs with Live Translation Enable real-time participation from multiple regions simultaneously through interactive virtual platforms with simultaneous interpretation. This reduces travel barriers while maintaining genuine dialogue across time zones, ensuring participation from nations with limited resources. 4. Sectoral "Governance Sprint" Workshops Convene 3-day intensive workshops (healthcare, finance, education, public administration, climate) where sector experts, regulators, civil society and technicians co-design governance frameworks. Products are concrete, tested approaches—not abstract principles. 5. Youth-Led Innovation Challenges Invite young technopreneurs and civil society activists to propose innovative governance solutions addressing AI's risks in their regions. Winners present at the main Dialogue, ensuring future-focused perspectives shape decisions and collective governance. 6. Peer Learning Networks Establish post-Dialogue communities of practice: SIDS peer learning on capacity-building; regulators sharing implementation experiences; civil society networks monitoring AI impacts. Dialogue becomes ongoing, not episodic. 7. Adaptive Feedback Mechanisms Use real-time polling, digital feedback tools and community surveys to assess whether dialogue is producing shared understanding or merely exposing entrenched positions. Adjust format mid-dialogue based on evidence. 8. Documented Outputs & Accessible Resources Ensure all dialogue outcomes are published in multiple languages, formats (written, video, infographic) and platforms accessible to practitioners globally, not just diplomats. Effectiveness requires moving from "talk about AI governance" to "collectively designing AI governance." Interactive, place-based formats that include all nations and communities produce better outcomes than passive listening—fulfilling the Global Dialogue's mission.

1. Mauritius's FAIR Guidelines Framework A non-binding, principles-based governance model emphasizing Fairness, Accountability, Inclusivity, and Responsibility. Rather than premature legislation, FAIR creates a stable reference point for policymakers, regulators, and courts to progressively introduce safeguards as capacity and evidence grow-enabling innovation while managing risks. 2. Sectoral Governance Approaches Tailored governance mechanisms for specific sectors based on risk and impact: Finance: AI risk assessments, algorithmic transparency requirements, bias testing in credit/lending decisions Healthcare: Clinical validation protocols for AI diagnostics, informed consent frameworks, audit trails for treatment recommendations Public Services: Algorithmic impact assessments, appeal mechanisms for AI-driven decisions, human oversight protocols Education: Bias audits in student assessment AI, transparency in learning algorithms, teacher-led implementation 3. Shared Governance Model Multi-stakeholder coordination translating principles into actionable mechanisms: Government: Policy development and legislative frameworks Private Sector: Responsible AI development and deployment standards Academia: Research, capacity building, and independent audits Civil Society: Rights protection and accountability monitoring 4. Adaptive Capacity Building Programs Progressive upskilling initiatives addressing governance capacity gaps: AI literacy programs for policymakers and regulators Technical training in algorithmic auditing and bias testing International partnerships for knowledge transfer and expertise access Regional peer-learning networks among SIDS and developing nations 5. National AI Governance Councils Inter-ministerial coordination bodies ensuring coherent governance across sectors, preventing regulatory fragmentation and enabling adaptive policy responses. 6. Algorithmic Accountability Platforms Transparent systems for: Registering high-impact AI applications Conducting algorithmic impact assessments Documenting bias testing and mitigation measures Enabling citizen grievance mechanisms and appeals 7. Data Protection & Privacy Frameworks Enhanced regulations addressing AI-specific risks: data minimization, algorithmic transparency, automated decision-making safeguards, individual rights (explainability, contestability). 8. International Standards Alignment Adopting relevant ISO/IEC and ITU standards while maintaining flexibility for local context-balancing interoperability with sovereignty.