White House Pushes AI Adoption Amid Agency Lag
The White House and OMB are pushing federal agencies to accelerate responsible AI adoption with new memos and ambitious targets. However, a new analysis suggests a wide gap remains between policy and execution, as agencies struggle with legacy IT, data silos, and compliance inertia.
The push for AI adoption is codified in OMB memos M-25-21 and M-25-22, which stem from a January 2025 executive order to remove barriers to AI leadership. A core directive mandates that federal agencies appoint a Chief AI Officer (CAIO) within 60 days and establish AI Governance Boards within 90 days to oversee strategy and implementation. For contractors, memo M-25-22 signals a shift in acquisition strategy, encouraging agencies to prioritize the purchase of U.S.-developed AI solutions. The new guidance also instructs procurement teams to build in protections against vendor lock-in, requiring provisions for data and model portability and knowledge transfers from incumbent contractors. Despite policy pressure, execution remains a primary hurdle, with a significant gap between ambition and implementation. The core challenge is that AI tools, which evolve weekly, are being forced through approval processes designed for predictable, slow-moving enterprise software, causing major delays. The workforce and data readiness problems are quantifiable. A recent survey showed 70% of federal leaders believe they need to upskill or hire new talent to implement AI effectively. Furthermore, 32% of leaders identified poor data quality as a top barrier, while 48% cited security and adversarial risk as the single biggest blocker to adoption. This push comes as AI use is already surging. According to a Government Accountability Office report, generative AI use cases across 11 major agencies increased ninefold between 2023 and 2024, jumping from 32 to 282. The total number of reported AI projects nearly doubled in the same period, from 571 to 1,110. The newly mandated Chief AI Officers are tasked with closing the implementation gap by coordinating AI use, managing risks, and promoting innovation. By December, these CAIOs must verify that "concrete safeguards" are in place for any AI system that could impact the rights or safety of Americans, or else cease its use
