This week saw major AI legal changes worldwide. In the U.S., California now requires AI-made images like TikTok filters to have hidden labels so people know they’re fake. Colorado made companies using AI for jobs or loans prove their systems aren’t unfair to minorities. Texas proposed a law to stop AI from creating social scores (like China’s system) but left a loophole for humans to do it.

The European Union started strict checks on high-risk AI systems, like robots that inspect car parts in Germany. The U.S. and China agreed to share an AI emergency hotline to prevent accidents and work together on medical AI standards.

Many companies are struggling with new rules. A survey found 60% of U.S. firms don’t have good systems to check if their AI follows laws. Startups like Credo AI sell tools to scan AI code for problems starting at $50,000 per year.

The FCC delayed new rules for stopping unwanted AI calls until 2026. A new Senate bill wants to let the FCC collect fines directly from companies breaking AI call laws.

Lawyers warn that AI legal trouble is rising. A New York company was sued because its AI hiring tool allegedly rejected older applicants, and a court forced them to share their AI’s training data.