Governments around the world are working hard to make new rules for AI agents - computer systems that can make decisions and take actions on their own. In Europe, leaders agreed on March 13 to make their AI rules simpler and clearer. They want to stop people from using AI to create fake sexual pictures or videos of real people without permission. Meanwhile, in the United States, there is a big problem: different states are making their own AI rules, and these rules sometimes contradict each other. The U.S. government's Department of Commerce released a report about this conflict to help create national rules that all states follow. Experts are also warning that companies need to be careful with AI agents because they can make mistakes that hurt people. By the end of 2026, there could be over 1,000 lawsuits against companies because their AI agents caused damage. To help solve these problems, a group of security experts created something called the OWASP Top 10 for Agentic Applications to teach companies how to keep AI agents safe. In the United States, more than 930 organizations sent feedback to scientists at NIST about how to make AI agents more secure. The main message from companies is clear: rules should be flexible and not too strict, so AI can continue to improve and help businesses grow. Across the whole world, there are now over 140 organizations working together on AI governance - deciding what rules are fair and helpful.