This weekly update covers major developments in legal and regulatory frameworks for agentic AI across the globe. Legal experts warn that most companies' old technology contracts do not cover the new risks created by AI agents. These outdated agreements were written when software needed humans to control it, but now AI agents make decisions on their own. When something goes wrong, companies often must pay for the damage, not the companies that built the AI.

In Europe, a big legal change is coming. The EU AI Act will start being enforced on August 2, 2026. This law has strict rules about how AI systems must be transparent and controlled. Companies operating in Europe must follow these new rules, and organizations are racing to get ready.

In the United States, different states are passing their own AI laws. California, Texas, and Colorado have passed laws about AI transparency and how AI should be governed. These laws went into effect or will soon. Meanwhile, the SEC (a government agency that oversees financial companies) is now checking if companies are lying about their AI capabilities.

Companies are finding new solutions to manage these risks. One approach is called "policy as code," which means writing rules that AI agents must follow automatically, like guardrails on a highway. Another important step is making sure contracts clearly spell out who is responsible when AI makes mistakes. Legal experts say companies need to update their contracts now, before problems happen.