A survey of 44 civil society organisations conducted by The Future Society found overwhelming consensus that AI governance must include legally binding prohibitions on certain high-risk AI systems and mandatory independent third-party audits of general-purpose AI. The 2025 Paris AI Action Summit saw these organisations push collectively for binding accountability mechanisms, AI workforce observatories, and structured public participation in AI governance decisions. The organisations explicitly reject voluntary frameworks as insufficient for managing AI risks at scale.
Published by: The Future Society (international AI governance think tank), based on a survey of 44 civil society organisations.
Key finding: Civil society organisations across the world have reached consensus that voluntary AI principles are insufficient. Their specific demands — legally binding prohibitions, mandatory independent audits, and public participation mechanisms — reflect standards applied to other high-risk industries.
Context: Civil society advocacy at the Paris AI Action Summit marks a shift from commentary to organised demand for enforceable rules. The breadth of the coalition, 44 organisations across regions, makes this a significant indicator of where non-governmental governance pressure is heading in 2026.