Nature and People Foundation

The Global Dialogue must create a structured, high-level convening process that brings together heads of states and governments, scientists, civil society, and responsible private-sector stakeholders to reduce strategic misunderstandings, broaden participation, and build practical pathways toward shared guardrails. The purpose and unique value of the Global Dialogue: • Focus on Governance. The Dialogue must remain focused on AI governance issues, rather than general discussions about the technology (other platforms exist for this, such as the AI for Good Summit). Its primary objective is to discuss transboundary issues that cannot be addressed at the national level: deepfakes, fraud, rogue AI agents. • Bridging Role. The UN must play a leading role in the AI Governance. The Dialogue serves as a platform that connects various regional and sectoral groups (BRICS+, SCO, G20), identifies gaps, and promotes convergence of approaches. Special attention my be paid to the statement by the Council of Heads of State of the SCO Member States on Further Deepening International Cooperation in the Field of AI (Tianjin, 2025), Statement by the BRICS Leaders on Global Governance in the Field of AI (Rio de Janeiro, 2025), Statement by the Council of Heads of State of the CIS on the Development of Cooperation in the Field of Civil AI (Moscow, 2024), Statement by the Council of the CSTO Parliamentary Assembly in Connection with the Development of Artificial Intelligence Technologies (Almaty, 2024). • Practical Focus. It is important not just to exchange opinions, but to develop practical tools and learn from the real-world experience of AI implementation in different countries and sectors. Systemic collaboration between all UN entities responsible for AI development is essential. This must be done at a systemic level to clearly understand where and who is responsible for what within the organisation's strategic planning and everyday work.

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Safe, secure and trustworthy AI

There must be a special focus on connectivity between AI and environmental challenges approached in an integrated, balanced manner. AI must be treated not only as a source of environmental pressure, nor only as a tool for solving environmental problems. Instead, it brings together environmental cost, technology, governance benefit, institutional response within one analytical frame. The global AI divide continues to widen, with AI resources and capabilities concentrated in individual countries and corporations. This concentration fundamentally limits the development of socio-culturally and linguistically contextual AI solutions, limiting AI's potential to accelerate progress toward collective development goals. Special attention must be given to BRICS+, the Global South, and multipolar governance. The comparative framework must include BRICS (notably, China, India, Russia), the EU, the USA. This is an opportunity to build the comparison around a shared set of practical questions, such as environmental disclosure requirements, renewable electricity arrangements, data-center constraints, and the degree of linkage between AI and environmental governance. Such approach must include governments, private sector, international organisations, civil society within one governance framework to clarify what each type of actor is particularly well placed to do, and how these roles can reinforce one another. The rapid development and deployment of powerful generative AI models has both environmental benefits and consequences, including increased demand for electricity and water consumption. Emerging technologies - are seen as strategic sectors that can simultaneously promote economic growth, energy security, environmental improvement. Breakthroughs in clean energy, electric vehicles, and smart infrastructure are transforming the way economies function and societies address environmental problems. This convergence of sustainable development, technological innovation, and global collaboration represents a triple transformation - decarbonisation of the economy, digitalisation of energy systems, and restructuring of international cooperation - that is transforming global efforts to combat climate change, prevent biodiversity loss, desertification and land degradation, water scarcity.

A special attention must be paid to the environmental agenda. This year is a super year for the UN environmental conventions: in August Mongolia will host the 17th session of the Conference of the Parties of the United Nations Convention to Combat Desertification (UNCCD), in October Armenia will host the 17th session of the Conference of the Parties of the Convention on Biological Diversity, in November Turkey will host the 31st session of the Conference of the Parties of the UN Framework Convention on Climate Change, and in December UAE will host the 2026 UN Water Conference. As of today, the only convention that presented its first report during the "AI for Good 2025" forum on this topic is UNFCCC: "Artificial Intelligence for Climate Action: Advancing Mitigation and Adaptation in Developing Countries" AI is increasingly recognized as a crucial tool for achieving the goals of the Convention on Biological Diversity (CBD), particularly in supporting the Kunming-Montreal Global Biodiversity Framework (GBF) aimed at halting and reversing biodiversity loss by 2030. The UNCCD has increasingly integrated AI into its global strategy to monitor land degradation and enhance drought resilience. Key initiatives include the development of AI-driven platforms for real-time monitoring and predictive analysis to support policymakers in making data-backed decisions for sustainable land management There is an urgent need to ensure and enhance coordination and joint action of the Rio Conventions and other relevant MEAs in this area.

Russia is the world's largest country, spanning over 17.1 million square kilometers (roughly 11% of Earth's land) across Europe and Asia. It is nearly twice the size of the next largest country (Canada) and features 11 time zones, diverse terrain, and roughly 144 million inhabitants. Forests cover 46% of the country's territory, and specially protected natural areas account for 13% of its total area. These areas are characterized by their considerable length, remoteness, and varied temperature and landscape characteristics. Over 13,000 people work in Russia's protected area system, and each ranger oversees 50,000 hectares of forest. For successful, long-term, and sustainable management and protection of these natural resources, the latest innovative technological solutions must be utilized. It is proposed to combine the work of environmental protection agencies, technology organizations, IT companies, and the expert community to improve the lives of citizens and preserve unique ecosystems. To preserve these unique ecosystems and improve citizen well-being, we propose a collaborative framework uniting environmental agencies, technology companies, and the expert community. Russia has also joined the energy race. The digital economy, including AI tools, blockchain, cloud computing, data storage and processing facilities is driving growing demand for electricity. According to some estimates, the energy consumption of data centers worldwide is already comparable to that of all heavy industry. Access to fuel and energy, the availability of networks and pipelines, and the installed capacity of power systems are playing a decisive and increasingly significant role.

There is a fundamental conviction: in a multipolar world, effective AI governance cannot be designed by one group of countries and exported to the rest. It must be co-created through inclusive dialogue that respects the sovereignty, priorities, and contributions of all stakeholders. The BRICS Leaders' Statement on the Global Governance of AI places special emphasis on equitable geographic representation in global AI governance — a principle this paper embodies in its authorship and analysis. During this process the role of AI Dialogue can't be overestimated as global cooperation is needed in governance of the new emerging technologies. In a context in which many discussions, initiatives and approaches related to AI governance are already under way at different levels, the Dialogue can help connect those efforts, identify practical lessons, and promote greater coherence, interoperability and mutual understanding across different processes, where feasible. It can help strengthen complementarities across the broader landscape and provide an inclusive UN platform for continued discussion and cooperation among Governments and relevant stakeholders.

Special attention my be paid to the international initiatives presented by the SCO, BRICS+, CIS and by the CSTO. The UN is paying particular attention to global standards and regulations for AI development, which will likely be reflected in the organization's post-2030 development goals. We are facing an institutional gap. The architecture for AI governance at the international level is itself in flux. The WSIS+20 review, concluded in December 2025, made the Internet Governance Forum a permanent forum of the United Nations. The UN General Assembly adopted Resolution A/RES/79/325 in August 2025, establishing the Independent International Scientific Panel on AI and the Global Dialogue on AI Governance. The inaugural Global Dialogue, co-facilitated by Estonia and El Salvador, will convene in Geneva in July 2026 alongside the AI for Good Global Summit. Separately, the World Artificial Intelligence Conference (WAIC) in Shanghai continues to serve as a key platform for international AI cooperation; at WAIC 2024, UNIDO inaugurated its Global Alliance on AI for Industry and Manufacturing Centre of Excellence. India AI Impact Summit stood out as the first major event of its scale held globally. It follows the lead of the AI Safety Summit held in the UK in 2023, the AI Summit in South Korea in 2024, and the AI Action Summit in France in 2025. The New Delhi Declaration on AI Impact was adopted at the India AI Impact Summit 2026. It was endorsed by a total of 88 countries and international organisations, reflecting a broad global consensus on leveraging AI for economic growth and social benefit. The declaration sets out a shared roadmap for AI governance and international cooperation. Therefore, environmental dimension of these emerging governance initiatives remains underdefined.

The AI Dialogue must be structured on four main pillars. Inclusiveness. The Dialogue must be actionable, enabling countries and communities to fully participate and benefit. Coordination. The Dialogue aims to be a unique UN platform that brings together existing initiatives, avoids duplication, and promotes convergence. Complementarity. The Dialogue and the Scientific Group are distinct, yet complementary and mutually reinforcing. The Scientific Group provides scientific evidence; the Dialogue provides Member States and stakeholders with the policy space to act on the findings. Outcome-oriented. This process aims to establish a robust foundation for long-term international cooperation on AI governance. The Global Dialogue must become a tool that will allow the Global South not only to receive technologies but also to participate in their creation.

Global Dialogue on AI must promote inclusion and support the voices of the Global South to be presented: from presence to co-authorship: • The Problem of "Imposed Decisions". Countries in the Global South often act as recipients of technologies and regulations, rather than co-authors. Governance frameworks need to be "globally informed" rather than "globally imposed." • Practical Inclusion. Participation should not be formal. Mechanisms are needed that allow local communities, including traditional institutions, to influence the agenda and present their governance models together with governments and corporations. • Infrastructure Gap. A key challenge for developing countries is the lack of access to computing power, quality data, and skills. Without addressing this issue, any discussion of regulation will remain elitist. • A "floor" rather than a "ceiling" of regulation. Instead of setting high standards that developing countries cannot meet (and therefore will ignore), it is proposed to establish a minimum global standard (governance floor) that ensures basic rights protection and allows for development. • Linguistic and cultural diversity. AI must reflect the real diversity of languages and cultures. Models trained primarily on English data reinforce digital colonialism. Investment in multilingual datasets and model evaluation with native speakers is needed.

The international community is establishing the institutional architecture for AI governance – from the UN Scientific Panel on AI to the BRICS+ AI Alliance to China's AI governance regulations. The environmental dimension of this architecture remains underdefined. Critically, this is being done in a context where the global AI landscape is no longer dominated by a single center of gravity. China leads the BRICS in AI development, with the realized economic potential of generative AI across BRICS nations projected to reach $600 billion by 2030. India, with the world's largest population and a rapidly growing AI ecosystem, hosted the first major AI governance summit in the Global South in February 2026. Russia, as the country with one of the most used languages on the web and a developed domestic AI ecosystem, has initiated the BRICS+ AI Alliance and contributes uniquely to multilingual AI development – a significant consideration given that approximately 90% of AI training data is in English, despite English being native to only about 18% of the world's population. The effective governance of AI's environmental impact cannot be designed from any single national or cultural vantage point. It requires the kind of inclusive, multistakeholder dialogue that the Internet Governance Forum was created to foster and that the Global Dialogue on AI Governance is mandated to advance.

In February 2025, at the AI Action Summit in Paris, France, UN Environment Program (UNEP), and the International Telecommunication Union (ITU) launched the Coalition for Environmentally Sustainable Artificial Intelligence, bringing together over 100 partners - including 37 technology companies, eleven countries, and five international organisations - for collaborative action on standardized measurement, life-cycle analysis frameworks, and prioritization of sustainable AI research. In June 2025, UNEP released its Sustainable Procurement Guidelines for Data Centers and Servers, the first international framework designed to help governments reduce the energy and water consumption of data center operations. These developments signal a growing institutional momentum - but also underscore that the environmental governance of AI is still in its formative stages, with significant gaps remaining in measurement, disclosure, and enforcement. UNEP's 2024 issue note Artificial Intelligence End-to-End explicitly called for comprehensive assessment of AI's environmental impact across the full lifecycle - from material extraction and hardware manufacturing through operation and disposal - noting that "governments are racing to develop national AI strategies but rarely do they take the environment and sustainability into account." ITU, through its Green Digital Action initiative, has launched a dedicated Sustainable AI working group and is developing technical standards for environmental sustainability in AI and data center operations. ITU reports examine AI and the environment alongside associated technical standards development, establishing the groundwork for internationally harmonized measurement frameworks. International Organization for Standardization (ISO) has published TR 20226:2025, which incorporates sustainability considerations across the full AI lifecycle, including computational load, resource use, carbon impacts, pollution, disposal, and operational characteristics. It is critical to use the accumulated best practices and international initiatives developed by the SCO, BRICS+, CIS and CSTO.