The AI issue at the 80th UN General Debate

By Javier Surasky


During its 79th regular session, the General Assembly adopted its first two resolutions on Artificial Intelligence—a topic that has become increasingly present in recent general debates in recent years.

What can we expect to be discussed on this issue in the 80th session? Which actors should we pay special attention to? What could become the central points of discussion? Let’s explore the landscape to assist those interested in the topic in following the speeches during the upcoming General Debate.

We believe three main issues will structure the discussion: models of AI governance, its environmental footprint, and the rapid widening of the digital divide between digitally central and peripheral countries.

Within those topics, different positions will emerge regarding: resource metrics (understood as concrete figures on electricity and water consumption by AI systems); financing of digital capacity development—with references to "Compute for All," "Green Gigafactories," "Digital Commons Facility," and "AI for Sustainable Development"—seeking to balance the burdens and opportunities between AI-leading countries and those falling behind; and AI governance modes and forums, with potential disagreements over whether to formally establish a UN Scientific Panel, a World AI Cooperation Organization, or to give a greater role to the ITU. We may also see debates on how best to articulate AI with the UN80 initiative, which has thus far placed excessive expectations on what AI can deliver as part of a broader UN reform.

A quick mapping of countries and positions allows us to foresee the prominence certain actors will acquire:

The European Union will advocate for the regulation of AI in such a way that gives states control over risks without halting progress, proposing a unified combination of “standards and resource allocation.”

In this regard, it will showcase to the international community its AI Act and the InvestAI fund, endowed with €200 billion, €20 billion of which will be allocated to building gigafactories (each hosting up to 100,000 chips to train AI models, aiming to increase Europe's computing capacity while reducing carbon emissions). This may lead to calls for transparency in water and energy use at data centers.

The United States will continue its “innovation first” approach—that is, AI regulation must not hinder or slow AI progress. Recall that in a strongly worded speech at the February AI Summit in Paris, Vice President JD Vance warned that “excessive regulation could kill a transformative sector” and refused to sign the declaration on “Inclusive and Sustainable AI.

Washington will likely avoid referencing energy resource consumption, and if it does, it may be to point an accusatory finger at China. It might also promote its cloud services as the fastest route for developing countries to “leapfrog” over the digital divide.

China will be the third major actor. It is a given that it will defend a model of AI governance centered on state sovereignty but cooperative in nature. Recall that the Global AI Governance Action Plan, launched by Premier Li Qiang in July 2025, suggests establishing a World AI Cooperation Organization within the UN system and promoting the use of open-source platforms. Both elements could be presented as opportunities for lagging countries to join the digital wave. It is also expected to make strong reference to Western export controls on AI-related chips.

India’s position may vary depending on how U.S.-imposed sanctions for its Russian oil purchases evolve in the coming weeks. It is almost sure that India will defend the general “AI for Development” perspective. As a precedent, at the Paris AI Summit, Prime Minister Modi called for “democratizing technology” and invoked G20 language on energy-efficient AI at scale for the Global South. India may support calls for increased concessional financing to develop green data center corridors, especially in South Asia and Africa.

Kenya is leading Africa’s demand for access to computing power. At a panel organized by the President of the General Assembly and held in Kenya this past July, the Group of Friends for Capacity Development in AI was established. Its position is that without preferential access to cloud services and the promotion of renewable energy, AI will deepen neocolonialism and dependency in the region. Kenya is expected to link demands for digital justice and climate justice, possibly providing data on how AI strains water resources in arid African areas.

Brazil will not only "open the game" as the first country to deliver its speech, but we can be sure it will connect AI governance to the climate agenda of COP30, which will be held this year in Belém do Pará. Brazil is likely to support establishing global standard rules on disclosing energy and water consumption for AI purposes. It may propose allocating part of the Loss and Damage Fund to the development of sustainable and resilient computing infrastructure. This position could be joined by the Small Island Developing States (AOSIS), particularly Barbados, calling for mandatory environmental impact assessments as part of the implementation measures of the Global Digital Compact.

The Nordic countries (Norway, Iceland, and Sweden) will aim to present themselves as examples of green AI development, offering their abundant water resources and low temperatures as incentives for the installation of low-emission AI clusters in their territories.

Iraq, speaking on behalf of the G77 + China, will once again call for greater technology transfers—including digital technologies—more financing for AI development, digital infrastructure, and human capacity. There is a strong likelihood that it will call for the creation of a global solidarity mechanism. The digital divide and data neocolonialism may also feature in its speech.

France and Spain will highlight their hosting of the AI Action Summit and the Fourth Conference on Financing for Development. While aligned with the EU, they may include elements suggesting the need for phased thresholds for computing and financial resource transfers, as well as for providing concessional loans for the development of "green AI." They may also advocate incorporating this issue into the agendas of multilateral development banks.

Taken together, these interventions reflect the contending positions around the issues we identified as the current core of debate, intertwining development, environmental stewardship, ethics, and international security—key components of a contested roadmap toward digital development.