This week brought major updates on how governments around the world plan to control AI agents - smart computer systems that can make decisions on their own.

The European Union is leading the way with its new AI Act, which puts strict rules on high-risk AI systems. These rules now apply to AI agents that work without human control. Companies must explain how their AI makes decisions and prove it won't cause harm.

The United Kingdom passed a new law called the Data (Use and Access) Act 2025. This law makes it easier for companies to use data for AI training, but it also creates new challenges for protecting people's privacy when AI agents handle personal information.

The United States is taking a different approach by letting each industry make its own AI rules. Healthcare, finance, and defense sectors are creating their own safety guidelines for AI agents.

China is focusing on government control of AI systems. They want to make sure AI agents follow national goals and are transparent about how they work.

Experts warn that AI agents create new problems that old laws didn't expect. These systems can make hundreds of small decisions very quickly, making it hard to track what went wrong if something bad happens. They can also learn and change over time, which means they might drift away from their original safe settings.

Companies are struggling with transparency requirements. Many AI agents work like "black boxes" - nobody can see how they make decisions. New laws require companies to explain their AI's choices, but this is very hard to do with complex agent systems.

The gaming industry faces special challenges when AI creates content in real-time. Europe's Digital Services Act requires quick removal of illegal content, but this becomes complicated when AI agents generate new material constantly.