This weekly update covers major legal and regulatory developments for agentic AI systems—artificial intelligence that can work independently, make decisions, and take actions without constant human approval.

The biggest news comes from Singapore, where the government announced the world's first governance framework specifically designed for agentic AI on January 22, 2026. The Model AI Governance Framework (MGF) was created by Singapore's technology authority to help organizations safely use AI agents. This framework focuses on four main areas: carefully checking risks before using AI agents, limiting what the agents can do, making sure humans stay in control, and teaching users how to work with these systems properly.

The framework requires organizations to set clear boundaries for their AI agents, such as limiting what information they can access and what decisions they can make without asking a human for approval. It also emphasizes that humans must stay responsible for what the AI agents do, and companies should continuously watch their agents' behavior to catch problems early.

Beyond Singapore, other countries are also focusing on how to regulate agentic AI safely. The United Kingdom published a report about data protection concerns with agentic AI, warning that organizations remain responsible for keeping personal information safe when using AI agents. Meanwhile, new United States state laws in Illinois, Texas, and Colorado are putting rules in place for how AI systems can be used in the workplace and for consumers.