From Evidence to Action: What Does the Global South Want from Global AI Governance?

02 de novembro de 2026

The global governance of artificial intelligence took on an even greater dimension over the course of 2026. In July, the United Nations released the Preliminary Report of the Independent International Scientific Panel on AI, the first global, independent scientific assessment dedicated to gathering evidence on the opportunities, risks and impacts of artificial intelligence. In December, the Data Privacy Global Conference will host the participatory workshop “MAP-AI: from Evidence to Action” to discuss some of this report’s findings and turn its evidence into concrete recommendations, from the perspectives of the Global South and, above all, of experiences in Latin America.

Created by the UN General Assembly, the Panel brings together 40 experts from different regions of the world, and its mission is to help ensure that international decisions on AI are informed by independent scientific evidence. The report starts from a fundamental question: how do we govern a technology that develops faster than our capacity to understand, assess and regulate it?

Its importance lies precisely in offering a common evidence base for a debate that, until now, has been marked by different perspectives, interests and levels of technical capacity. The document identifies seven major areas, including AI science and development; social applications; economic impacts; security and the environment; human rights, information and democracy; child autonomy and safety; and management, governance and reliability.

The report itself identifies a significant inequality in countries’ capacity to take part in AI governance. Computing infrastructure, data, technical expertise and assessment capacity are concentrated in the countries and companies that develop the most advanced systems. As a result, many countries remain dependent on technologies they cannot fully build, inspect, audit or adapt to their local contexts, an issue that is particularly relevant for the Global South.

AI systems are developed mostly on the basis of data, infrastructures, languages and cultural assumptions that do not represent all societies equally. The report draws attention, for example, to challenges related to the availability of data in different languages and to the need to adapt systems to local contexts.

At the same time, the report recognizes that global AI governance is still fragmented. There are dozens of regulatory and non-regulatory instruments, but they remain poorly coordinated and are often concentrated in the corporate sector. The document points to the need to combine different instruments, including binding legislation, sectoral regulation, technical standards, voluntary commitments and other forms of soft law.

It is precisely in this space between evidence and governance that our workshop aims to act.

At the Data Privacy Global Conference, members of the Global South Alliance (GSA) and the MAP-AI project will lead a participatory workshop dedicated to discussing two of the report’s central themes from the standpoint of Latin American experiences: human rights, information and democracy; and management, governance and reliability.

The Global South Alliance brings together civil society organizations from the Global South that work on research and advocacy on digital policy and international governance processes. Its members, like those of the MAP-AI project, have taken part in debates and processes within the United Nations system, including discussions related to the Global Digital Compact and the new Global Dialogue on AI Governance.

The Global South Alliance itself has argued that international AI governance processes should address structural inequalities, broaden the meaningful participation of Global South communities and strengthen mechanisms related to human rights, data governance and public interest infrastructure.

The workshop will begin with an overview of UN processes related to global AI governance and of the work of the Global South Alliance and MAP-AI in these spaces. Representatives of these networks will then present the two selected themes from the report. Participants will then be divided into groups to discuss concrete risks and opportunities and, above all, to think about what global AI governance should include.

The Panel’s next annual report is expected to feed into the second Global Dialogue, scheduled for May 2027 in New York. In other words, taking part in this conversation also means taking part in shaping what may be brought to the next global governance spaces.

Date: December 8, 2026
Time: 2:00–3:30 p.m.
Format: participatory workshop