Countries around the world are updating rules for AI agents to make them safer and fairer. In the United States, states like California and Colorado passed new laws requiring companies to be clear when they use AI. For example, California’s SB-942 law says AI-made content must have hidden labels showing it was created by a computer. Colorado’s AI Act will force businesses to check for bias in AI systems used for jobs or loans starting in 2026.

The European Union is leading global efforts with its AI Act, which sorts AI tools by risk level. High-risk AI, like those used in hiring, must pass strict tests and be transparent. Meanwhile, China released rules to control how generative AI services are used, focusing on stopping fake news.

In the U.S. federal government, President Trump replaced old AI rules with new ones aiming to boost innovation. His order tells agencies to support AI growth while cutting rules seen as too strict. Lawyers predict more states will copy the EU’s risk-based approach but might struggle to balance safety and innovation.

Companies worldwide are preparing for these changes by adding AI audits and training. Experts warn that failing to follow new laws could lead to big fines or lawsuits.