6.0 KiB
IGF 2026 WS #290 Building a global AI accountability infrastructure
Description
International human rights law already provides an important normative baseline for AI governance. Across the UN system, multiple human rights mechanisms and mandates have consistently affirmed that technologies incompatible with international human rights law should not be developed, deployed or used. Yet translating this normative consensus into effective accountability remains an unresolved challenge. This session explores how global AI governance can move from shared principles to operational accountability. Building upon the work already done by human rights mechanisms and the Scientific Panel’s report, it will focus on the institutions, evidence systems and coordination mechanisms required to identify harms, monitor compliance and enable accountability across borders. The discussion starts from a simple observation: AI production, deployment and impacts are transnational, while accountability remains fragmented. AI systems rely on global supply chains whose environmental, labour and human rights impacts are distributed across jurisdictions. Many harms first become visible through affected communities, while existing international governance processes often lack systematic ways to collect, validate and incorporate this evidence into decision-making. Building on existing international human rights standards and the momentum generated by the first UN Global Dialogue on AI Governance, participants from the AI Scientific Panel, governments, international organizations and civil society will discuss the institutional architecture needed to operationalize accountability. The workshop will examine practical building blocks - including independent evidence generation, pathways for documenting harms, coordination among international institutions, and mechanisms connecting local experiences to global governance processes - while identifying priorities for the intersessional period leading to the second Global Dialogue.
Policy Question(s)
A. What institutional mechanisms are needed to transform international AI governance principles into effective systems of accountability across borders? B. How can existing international human rights and multilateral mechanisms, including the UPR, treaty bodies, specialised UN agencies and the Scientific Panel on AI, be integrated into a coherent global AI accountability infrastructure? C. How can local evidence of AI harms be systematically incorporated into global AI governance processes?
Expected Outcomes
The workshop will produce a set of recommendations on priority building blocks for a global AI accountability infrastructure, reflecting the multistakeholder dialogue held at the IGF. The recommendations will focus on operational measures capable of strengthening implementation of existing international human rights standards in AI governance. Positioned between the first and second UN Global Dialogues on AI Governance, the session treats the IGF as a key intersessional space to advance concrete proposals before negotiations resume in 2027. The recommendations will be shared with relevant UN processes and stakeholders involved in the Global Dialogue. Participants will leave with practical ideas on how existing international institutions can be connected into a coherent global accountability infrastructure.
Reference Document
Reference Document
Format
Classroom Duration (minutes): 90 Format description: A 90-minute classroom format provides sufficient time for a structured exchange among speakers representing different stakeholder groups and areas of expertise, followed by moderated discussion with participants. The session addresses a complex institutional challenge that cannot be meaningfully explored through short presentations alone. The classroom setting supports a focused policy dialogue while allowing participants to engage with speakers throughout the discussion, helping to test ideas, identify areas of convergence and explore different institutional approaches. The extended duration is particularly important given the breadth of mechanisms under discussion (from UN human rights bodies and specialised agencies to emerging AI governance institutions) and the opportunity to identify practical pathways for strengthening global AI accountability during the intersessional period before the second UN Global Dialogue on AI Governance.
Hybrid Format
Hybrid Format: The session will be designed to ensure that onsite and online participants can engage on equal footing throughout the discussion. The moderator will actively alternate between questions and comments from the room and the online platform, while a designated online moderator will monitor the virtual discussion and relay interventions in real time. The discussion will include dedicated moments for questions and comments following each thematic segment, allowing participants to shape the conversation as it develops. This approach encourages dialogue across stakeholder groups and regions while ensuring that both onsite and remote participants can meaningfully contribute to identifying priorities for strengthening global AI accountability. No additional online platforms are planned beyond the hybrid participation tools provided by the IGF Secretariat.
Organizers
Organizer 1: Marina Meira, Derechos Digitales Organizer 2: Souza Michel , OHCHR
Speakers
Speaker 1: Paloma Lara-Castro, Civil Society, Latin American and Caribbean Group (GRULAC) Speaker 2: Souza Michel , Intergovernmental Organization, Latin American and Caribbean Group (GRULAC) Speaker 3: Nighat Dad, Civil Society, Asia-Pacific Group
Moderator
Nighat Dad, Civil Society, Asia-Pacific Group
Online Moderator
Alithu Bazan Talavera, Civil Society, Latin American and Caribbean Group (GRULAC)
Rapporteur
Marina Meira, Civil Society, Latin American and Caribbean Group (GRULAC)
Participant Benefits
What will participants gain from attending this session?
SDGs
5.c 8.8 9.1 10.3 10.5 16.3 16.6 16.7 16.b 17.14