Trump administration weighs pre-release checks after AI deregulation

In short: The Trump administration is discussing oversight of AI models before they are released, even though its recent policy actions have focused on reducing federal rules.

What happened

The New York Times reports that the Trump administration is now discussing whether to impose oversight on some AI models before they are made publicly available. Pre-release oversight would mean some kind of review before an AI system can be shared widely, similar to checking a product before it goes on store shelves.

So far, the administration’s record has pointed the other way. On January 23, 2025, it issued an executive order called “Removing Barriers to American Leadership in Artificial Intelligence.” That order revoked the Biden administration’s October 2023 AI order, including requirements for safety “red-teaming” (stress testing a system by trying to make it fail), cybersecurity steps, and threat evaluations for some high-risk models.

Later, on December 11, 2025, the White House signed Executive Order 14365, which pushed for a national AI policy that could override state-by-state rules. On March 20, 2026, the administration released a National Policy Framework for Artificial Intelligence. It recommended Congress block state AI laws that “impose undue burdens,” rely more on existing agencies and industry standards, and avoid creating new federal rulemaking bodies. It also argued against penalties for developers if other people misuse their tools.

Why it matters

If the government moves toward pre-release checks, it could affect how quickly new AI tools reach the public and what safety steps companies must take. It could also change the balance between federal rules and state rules, which can shape what protections consumers get and what developers are allowed to build.

Source: NYTimes