Five civil society coalitions publish a read-out report on the first UN Global Dialogue on AI Governance

Five civil society coalitions urge a more inclusive, rights-based and accountable second UN Global Dialogue on AI Governance.

31 Jul 2026Civil societyGlobal Network Initiative ↗

Summary

A read-out report dated 31 July 2026 reflects on the first UN Global Dialogue on AI Governance. Five civil society coalitions wrote it together: MAP-AI, the Global South Alliance, the Global Digital Justice Forum, the Global Digital Rights Coalition, and the civil society participants in the Gender in Digital Coalition. It offers recommendations for making later sessions more inclusive and rights-based, noting gaps to close before a second Dialogue.

According to the report, women and Indigenous Peoples were insufficiently represented, human rights discussion was uneven across themes, and direct exchange between governments and other stakeholders was limited. Its recommendations follow the Dialogue's four thematic clusters, covering opportunities and implications, bridging AI divides, safe and trustworthy AI, and human rights, plus procedural proposals.

On substance, the coalitions want the environmental, labour and social costs of AI treated as core governance matters, with mandatory impact assessments along the value chain and a precautionary baseline. They note that 2.2 billion people are offline and largely absent from training data, and ask that consent-based data on under-documented groups count as a core digital foundation. They call for standardized, human rights-based benchmarks, social audits, gender impact assessments, and zero tolerance for technology-facilitated gender-based violence. On rights, they want the Chair's summary to affirm that international law already applies to AI, to promote due diligence, to address liability gaps, and to call for prohibitions or moratoria on unacceptable uses. The illustrative list includes biometric mass surveillance, social scoring, AI systems that enable lethal force, and AI systems that target decisions where meaningful human control is missing.

Procedural ideas include having the next Co-Chairs attend the Internet Governance Forum, resourcing the UN human rights office, helping Global South participants obtain visas and giving civil society speaking roles.