This weekly update covers important developments in agentic AI regulation across the globe. The United Kingdom announced new regulatory sandboxes for AI companies and invited feedback on how rules should handle agentic AI, which is a smarter type of artificial intelligence that can make decisions on its own. The European Union continued rolling out its AI Act, with new rules for general-purpose AI models becoming active in August 2025, and the full law taking effect in August 2026. Meanwhile, Singapore released draft guidelines specifically for agentic AI, inviting feedback from businesses through December 2025. These moves show that governments worldwide are racing to create rules for this new technology before it becomes too widespread. The main challenge is finding the right balance: rules need to be strong enough to keep people safe, but not so strict that they stop companies from creating new and helpful AI tools. Regulators are trying different approaches—some countries prefer government control, others want businesses to follow voluntary guidelines, and some use a mix of both.