After months of debate over whether powerful AI models should need government approval before release, the White House chose a lighter path. President Donald Trump's June 2 order, "Promoting Advanced Artificial Intelligence Innovation and Security," focuses mainly on government cyber defense rather than premarket model review.
The order gives agencies fast deadlines: CISA must issue binding operational directives and create an AI-driven defense program within 30 days; Treasury is to build an AI cybersecurity information exchange with industry; defense systems are to be hardened; and within 60 days Treasury, defense and homeland security agencies must produce AI model benchmarking procedures.
The most important sentence says nothing in the section authorizes a mandatory governmental licensing, preclearance or permitting requirement. That language shuts down, at least for now, a more FDA-like approval regime that had worried technology companies and investors.
The result is a compromise: standards and federal preparation, but no compulsory release gate. Critics will argue that voluntary review has no teeth, especially for actors that do not care about reputation. The next test is whether the 60-day benchmarks become a de facto requirement through procurement, infrastructure rules or state-level regulation.
Sources：CocoLoop、Trump's new AI safety order seeks voluntary review of new models (NPR); Promoting Advanced Artificial Intelligence Innovation and Security (The White House); White House Considers AI Vetting, Sparks Tech Industry Panic (The Hill)