States move ahead with AI rules despite Trump order to stop them

In short: U.S. states are putting new AI rules in place in 2025 and 2026, even after a White House order signaled it wants to block state laws.

What's going on

Several states are moving forward with laws meant to protect consumers and reduce harm from AI systems, including generative AI (tools that create text, images, or video). The push is happening even though President Trump signed an executive order in December 2025 that pointed to federal plans to disrupt or override state AI rules.

Utah already has a law in effect. Utah SB 226, effective May 7, 2025, focuses on AI used in consumer interactions, especially in regulated areas like finance and health. In some cases, companies must clearly tell people they are interacting with AI, and the Utah attorney general can seek penalties of up to $5,000 per violation.

California has multiple AI laws scheduled to take effect on January 1, 2026. One of the biggest is AB 2013, which requires developers of generative AI to publish summaries of the data used to train their systems. Think of it like an ingredients label for an AI model, listing the types and sources of data. California also passed a separate transparency law, AB 853 and SB 942, that later requires tools to help detect AI-made content and to add hidden markers that help trace where content came from.

Other states are also considering rules, including Texas, which is advancing a broader AI governance bill for 2026.

What to watch

The big open question is whether the federal government can and will block state laws, and what that means for businesses and consumers. If states keep moving, companies may face a patchwork of rules, like driving through different states with different speed limits.

Source: NYTimes