White House Releases New Guidance Expanding Federal AI Oversight
The White House set new requirements for neutral AI in federal contracts while pushing for a unified national AI policy.
The Office of Management and Budget provided new guidance for federal agencies Thursday on how to procure “unbiased” artificial intelligence, a followup to President Donald Trump’s July executive order.
The memorandum requires agencies to apply two core principles to AI systems they procure: truth-seeking and ideological neutrality. AI must provide accurate, evidence-based responses and clearly acknowledge uncertainty where information is incomplete or contradictory. It must also operate as a nonpartisan tool, avoiding the intentional encoding of political or ideological judgments, unless explicitly prompted or readily accessible to the end user, according to the memo.
Federal agencies will be required to integrate these principles into new procurement contracts and modify existing contracts where possible. The memo also directs federal agencies to require documentation from AI vendors that includes model and system cards detailing the training process, risk mitigation strategies and evaluation scores on benchmarks assessing accuracy and bias. Vendors must also supply acceptable use policies, user guides and mechanisms for end-user feedback to ensure AI outputs meet the federal guidelines.
Agencies are expected to update policies by March 11, 2026, to ensure compliance across all federal AI contracts. The memo also establishes a sunset provision, meaning the guidance will expire in two years unless extended by the OMB director.
White House Pushes for National AI Framework
OMB’s memo coincided with President Trump’s executive order, which seeks to limit the authority of states to regulate AI and establish a unified national AI policy.
“We remain in the earliest days of this technological revolution and are in a race with adversaries for supremacy within it,” Trump said in the order. “To win, United States AI companies must be free to innovate without cumbersome regulation. But excessive State regulation thwarts this imperative. … State-by-State regulation by definition creates a patchwork of 50 different regulatory regimes that makes compliance more challenging, particularly for start-ups.”
The order directs the attorney general to establish an AI Litigation Task Force to challenge state AI laws inconsistent with the order, specifically laws that have “cumbersome regulation” or “imbed ideological bias.” The administration argues that diverging state standards threaten U.S. competitiveness and could slow progress against global rivals, particularly China.
The order also directs the Special Advisor for AI and Crypto David Sacks to craft legislative recommendations for Congress that would establish a national standard and formally preempt state laws. Sacks said more than 1,000 AI-related bills are advancing through state legislatures. Allowing them to stand would leave the industry navigating “50 states running in 50 different directions,” he said Thursday during the Oval Office order signing.
This is a carousel with manually rotating slides. Use Next and Previous buttons to navigate or jump to a slide with the slide dots
-
DOL’s AI-Ready Initiative Gains Early Traction with Nationwide Rollout
A new seven-day AI course from the Labor Department uses mobile learning to boost AI literacy nationwide.
3m read -
Building the Military Health System’s AI Ecosystem
DHA and Red Hat discuss building a secure AI ecosystem, strengthening data foundations and ensuring safe AI use in military health care.
31m watch -
White House AI Framework Revives Early Internet Governance Debates
Trump administration’s National AI Legislative Framework echoes Web 1.0 tensions over liability, innovation and federal preemption.
4m read -
Marine Warfighting Lab Turns Lessons from Ukraine into Future Strategy
The Marine Corps Warfighting Laboratory uses wargaming and simulation to analyze threats, test capabilities and prepare for future conflicts.
12m listen