Singapore Unveils the First Global Governance Framework for Agentic AI
On January 22, 2026, Singapore introduced the Model AI Governance Framework for Agentic AI (MGF) at the World Economic Forum 2026, becoming the first country to establish a governance framework specifically dedicated to agentic AI.
The framework targets AI systems capable of reasoning, planning, and executing actions autonomously on behalf of humans. Although it does not create binding legal obligations, the MGF sends a strong signal regarding Singapore’s regulatory direction and provides operational best practices for organizations developing or deploying such technologies.
The MGF builds upon Singapore’s existing AI governance initiatives, including the 2019 Model AI Governance Framework, the AI Verify testing framework, and the Global AI Assurance Pilot launched in 2025. What distinguishes this new framework is its specific focus on the unique risks posed by agentic AI, such as unauthorized actions, autonomous errors, biased decision-making, data leaks, and disruptions to connected systems.
The framework is structured around four key pillars:
- Early risk assessment and risk bounding by design
- Clear human accountability
- Technical controls throughout the system lifecycle
- End-user responsibility and transparency
Through this approach, Singapore aims to enable the responsible adoption of increasingly autonomous AI systems while maintaining trust among users and organizations.
Source
Hogan Lovells – Singapore launches first global agentic AI governance framework