EU and US states tighten AI transparency rules as US federal law lags

In short: AI regulation is moving faster in the EU and in US states like California than at the US federal level, and the rules are getting more specific about transparency and labeling.

What's going on

The global rulebook for AI is still uneven, but it is getting stricter in key places. In the European Union, new transparency duties under the EU AI Act took effect on August 2, 2026. Transparency means AI providers must be clearer about what their systems do and how they are used, like putting a label on a product so people can tell what they are looking at.

At the same time, the European Commission can now enforce rules for providers of general-purpose AI models, which are broad AI systems that can be used for many tasks. The Commission can request information, order fixes, or even require a product to be pulled from the market. Fines can reach €15 million or 3% of worldwide annual revenue.

In the United States, the biggest moves are happening state by state. California’s AI Transparency Act also became operative on August 2, 2026. It applies to certain generative AI services with more than 1 million monthly users accessible in California and requires tools to verify provenance, meaning a way to check where AI-made images, video, or audio came from (like a receipt or watermark that helps trace a file’s origin). On September 18, 2026, Governor Gavin Newsom issued an executive order to speed up independent oversight and explore an AI “kill switch” idea for advanced systems.

What to watch

Congress faces growing pressure to pass a comprehensive national AI law, but none has passed yet. Until that happens, people and companies should expect a patchwork of rules, with clearer labels, tracking, and safety requirements appearing first in places like the EU and California.

Source: NYTimes