White House Halts Public AI Testing Reports Amid Security Push

The Trump administration has directed a federal AI testing unit to suspend publication of its public model assessments, marking a shift toward greater White House control over artificial intelligence evaluation as security considerations take priority.

Administration officials, including National Cyber Director Sean Cairncross, have instructed the Center for AI Standards and Innovation to stop issuing public reports while a recent executive order is implemented, according to people familiar with the matter. The Wall Street Journal first reported the development.

The directive comes after President Trump signed an executive order last week that represented a policy victory for Cairncross and Treasury Secretary Scott Bessent, both of whom have advocated for national security to play a more prominent role in how the government evaluates AI models.

Why it matters

The move signals a fundamental change in how the U.S. government approaches AI oversight. By halting public assessments from a unit that had earned praise from AI developers, the administration is prioritizing classified security reviews over transparency. This shift could affect how AI companies develop and deploy models, potentially slowing public understanding of AI capabilities and risks while strengthening government control over which technologies are deemed safe for deployment in sensitive contexts.

Testing unit's uncertain future

The Center for AI Standards and Innovation had been conducting evaluations of AI models and sharing findings publicly, work that won recognition from developers in the AI industry. The suspension of its public-facing activities throws the unit's future role into question, though it remains unclear whether the center will continue operating in a classified capacity or face broader restructuring.

The administration's approach reflects growing concerns within the government about the national security implications of advanced AI systems, particularly as models become more capable and potentially applicable to sensitive domains including defense, intelligence, and critical infrastructure.

Policy shift toward security

The executive order that precipitated the reporting halt marks a clear policy direction: security considerations will now take precedence in federal AI evaluation processes. This represents a departure from earlier approaches that emphasized public transparency and industry collaboration in AI safety assessment.

Cairncross, who leads the Office of the National Cyber Director, has been a key advocate for this security-focused stance. His office plays a central role in coordinating federal cybersecurity policy and has increasingly focused on AI-related risks.

Details of the executive order's specific provisions and implementation timeline were not disclosed in the report. The administration has not publicly commented on how long the publication suspension will remain in effect or what criteria will determine when or if public reporting resumes.

The Wall Street Journal reported these details on June 9, 2026.