How shifting governance structures, surging AI spending, and federal-state tensions reshape the public administrator's role
Federal AI spending jumped 966 percent between 2024 and 2026, reaching $7.2 billion. This explosive growth mirrors a deliberate policy pivot from cautious governance to accelerated execution. The Biden administration's Chief AI Officers Council, with 17 members, emphasized interagency coordination and risk management. By June 2026, the Trump-era council had slimmed to nine members and prioritized competitiveness and rapid adoption. For agencies, the new posture scrambles familiar compliance rhythms. Speed now competes directly with accountability, leaving public administrators to reconcile innovation mandates with the durable demands of equity, transparency, and public trust. Understanding this tension is essential for anyone studying public administration versus public policy, as both fields now intersect directly with how AI governance is designed and executed.
Federal AI policy has shifted from cautious governance to aggressive execution in just a few years. Public administrators who understand the key milestones, including executive orders, new oversight bodies, and binding guidance, can better anticipate where resources and regulatory pressure will be directed next. This timeline traces the major actions that have shaped the current landscape.
In October 2023, President Biden signed Executive Order 14110, the most sweeping federal action on AI to that point. The order set out to manage AI risks while promoting innovation, embedded in a whole-of-government approach. It directed the National Institute of Standards and Technology (NIST) to finalize a voluntary AI Risk Management Framework (AI RMF) as the government's foundational safety standard, a tool that would later be referenced in procurement guidelines, agency plans, and legislative proposals.
The order also required each major federal agency to appoint a Chief AI Officer (CAIO) and to publish an internal AI use-case inventory, making the government's growing reliance on algorithmic systems more transparent than ever before.
Building on the EO, the White House established the Chief AI Officers Council (CAIOC) in early 2024. With 17 members drawn from across the largest civilian and national security agencies, the Council was designed to coordinate risk management, share best practices, and harmonize AI implementation standards. Its creation signaled that AI oversight would not be left to individual agencies alone.
Simultaneously, the Office of Management and Budget (OMB) released binding guidance requiring agencies to designate CAIOs, complete AI use-case inventories, and apply the NIST AI RMF to high-risk systems. The OMB memo effectively turned the voluntary framework into a compliance yardstick for the executive branch. At this stage, the emphasis was squarely on safety, equity, and intergovernmental relations public administration.
The return of the Trump administration brought a rapid pivot. Early 2025 executive orders rescinded several provisions of EO 14110, replacing the risk-first posture with a mandate for speed and global competitiveness. Agencies were instructed to cut regulatory barriers, accelerate AI procurement, and prioritize adoption over deliberation. By June 2026, the CAIOC had been streamlined to nine members, reflecting a narrower, more execution-oriented mission.
Federal spending mirrors this urgency. Brookings researchers Denford, Dawson, and Desouza (June 2026) found that AI-related contract spending rose 966 percent from $355 million in 2024 to $7.2 billion in 2026, with potential award value surging to $91.8 billion.1 The number of agencies with AI contracts grew from 17 in 2022 to 28 in 2026, and the Department of Defense alone accounts for $90 billion in AI contract value. Total federal AI contracts jumped from 472 in 2022 to 1,743 in 2026.1
Congress has yet to pass a comprehensive AI law, but multiple bills introduced since 2022 have focused on algorithmic transparency, procurement guardrails, and workforce training. As of mid-2026, the most-watched proposals include measures to preempt state-level AI laws and to codify aspects of the NIST AI RMF. For public administrators, the legislative landscape remains fluid, and the gap between what Executive Orders require and what Congress may ultimately mandate continues to create uncertainty in long-term planning.
What changed when the Chief AI Officers Council transitioned from the Biden administration's risk-management focus to the Trump administration's emphasis on execution and competitiveness?
In early 2024, President Biden established the Chief AI Officers Council (CAIOC) to coordinate how federal agencies adopt and oversee artificial intelligence. The council's initial mandate centered on mitigating risks: ensuring AI systems were safe, trustworthy, and aligned with equity concerns. With 17 members drawn from major departments and agencies, the CAIOC served as a cross-cutting forum for sharing best practices around testing, auditing, and impact assessments. Publicly, the council issued guidance on bias mitigation, data privacy, and algorithmic accountability, reflecting a deliberate approach that prioritized safeguards before scaling up deployment.
By mid-2026, the council had been reconfigured. As reported by the Brookings Institution,1 membership had been trimmed to 9 officials, and the group's focus shifted toward accelerating AI adoption and maintaining global competitiveness. Where the Biden-era council operated as a collaborative governance body, the Trump-era version functions more like a task force for removing barriers to rapid implementation. Federal AI spending surged 966% between 2024 and 2026, with contract counts tripling over the same period, putting immense pressure on agencies to move from pilot programs to full-scale operational use. The council's activities now concentrate on procurement streamlining, workforce upskilling, and interagency data-sharing agreements to speed deployment.
This shift is not merely rhetorical. Agency leaders now face compressed timelines to integrate AI into mission-critical functions while still maintaining accountability. The governance scaffolding built under Biden, such as risk frameworks and bias audits, remains largely in place but is being reinterpreted through an execution lens. For practitioners, the operational question becomes how to preserve rigorous oversight within an environment that rewards speed. Key signals to track include new OMB guidance on AI procurement, updates to the Federal Register, and any formal directives from the council itself, which may or may not be publicly posted. Nonpartisan outlets like Brookings and CSIS continue to offer independent analysis of these developments, helping administrators navigate the evolving expectations without losing sight of equity and safety requirements.
Understanding this transition is essential for public administration jobs and policymakers who must now operate under a framework that balances two sometimes conflicting imperatives: risk avoidance and rapid innovation.
The scale of federal AI investment has transformed dramatically. Over two years, spending, contracts, and agency participation have grown by orders of magnitude, reshaping procurement and oversight demands for public administrators.
Federal agencies in 2026 occupy two distinct lanes when it comes to AI implementation: those with mature, deeply integrated programs supported by multi-billion-dollar contracts, and those still navigating pilot projects or initial governance frameworks. While the Department of Defense commands the largest share of federal AI spending, smaller civilian agencies are carving out niche applications that often fly under the radar. Public administrators looking to understand the implementation landscape must know where to look and how to interpret what they find.
Most major agencies now publish information about their AI activities on public .gov websites. Press releases, blog posts, and annual performance reports often describe new tools or outcomes. Many agencies also maintain online inventories of AI use cases, as encouraged by executive order and OMB policy. These inventories typically list the purpose of each system, its development stage, and the data it relies on. For example, health-focused agencies have disclosed applications for processing medical claims or forecasting public health risks, while veterans affairs offices have shared details on tools that streamline benefit determinations. Public administrators should review these inventories not just as lists, but as signals of where an agency is investing technical talent and infrastructure.
Two government-wide resources provide a broader view. First, the Government Accountability Office issues technology assessments and audits that sometimes profile agency AI efforts, highlighting both successes and recurring problems. Second, OMB publishes a consolidated inventory of federal AI use cases, compiled from agency submissions under the AI in Government Act. These reports are useful for identifying cross-cutting trends: common applications like process automation, fraud detection, and document analysis appear across many departments, while more specialized tools remain unique to a single agency. Administrators should note that centralized reports often lag real-time deployments by months, but they offer a baseline for comparing maturity across agencies.
For a more current picture, public administrators can turn to federal procurement data. Searching contract award databases reveals which vendors are delivering AI solutions, the scope of work, and the contracting vehicle used. This information can clarify whether an agency is building custom models in-house, buying commercial software, or adopting a hybrid approach. Industry associations like ACT-IAC and the Partnership for Public Service also publish case studies and host forums where agency leaders discuss implementation lessons, making them valuable for understanding the practical realities behind the contract language.
Implementation varies widely: some agencies have embedded AI into legacy workflows, while others are still standing up governance boards. Applying federal administration best practices can help administrators move beyond flashy press announcements and instead look for evidence of user testing, impact metrics, and plans for ongoing monitoring. Not every pilot results in a productive system, and public reports sometimes emphasize early wins over long-term sustainability. For administrators, the key is to triangulate across multiple sources: agency self-reports, oversight documents, and procurement records together form a more complete view of who is doing what in federal AI implementation this year.
Public administrators find themselves caught between two competing regulatory impulses: the federal government's push for a unified, innovation-friendly AI framework and individual states' efforts to establish their own, often stricter, governance models. As Colorado and California have already enacted AI laws, federal agencies are now exploring tools, from litigation to grant conditioning, that could override those state rules. For anyone managing AI programs across jurisdictions, the immediate question is which compliance regime holds sway when mandates conflict, and how to plan amid the uncertainty.
In December 2025, Executive Order 14365 introduced a formal mechanism for challenging state AI laws.1 It directed the Secretary of Commerce to evaluate state laws that might "alter truthful outputs" or impose "onerous regulation," with a report due by March 2026.2 The order also created a new AI Litigation Task Force within the Department of Justice, giving it 30 days to stand up.1 The task force was instructed to examine state laws under legal theories like Commerce Clause preemption, but the order itself does not automatically preempt any state law.3
As of mid-2026, however, the task force has not filed any public lawsuits.4 It has issued an internal framework document, but no state law has been formally challenged in court.5 The FTC and FCC also had deadlines to propose preemptive policies, yet those remain in process.1 This means the federal preemption threat is real but not yet active, leaving agencies operating in a legal gray zone.
Two state laws are widely seen as the most likely targets: the Colorado AI Act, effective June 30, 2026, and the California Transparency in Frontier Artificial Intelligence Act, effective January 1, 2026.2 Colorado's law imposes risk-based requirements for high-risk AI systems, while California's focuses on transparency for foundational models. Neither has been formally challenged, but the task force's early evaluations have reportedly flagged both because of their potential to conflict with federal goals of rapid AI adoption. For public administrators overseeing federal state partnership accountability tools, this fragmented landscape demands a proactive compliance posture rather than a wait-and-see approach.
In the short term, the safest path is to prepare your AI governance processes to adapt quickly. Document how your agency meets both federal and state standards, and maintain an open dialogue with legal counsel about evolving preemption risks.
How can public administrators manage AI compliance when formal mandates are still evolving and agency resources are thin? The answer lies in anchoring practices to the NIST AI Risk Management Framework (AI RMF), which has become the de facto standard across federal agencies and their contractors1, even as it remains voluntary under current executive orders.2 While the framework is strongly recommended and widely referenced, no unified federal compliance metrics exist yet3; the framework itself is under revision, with a critical infrastructure profile released in concept note form in April 2026.4 The Department of Treasury, for example, has mapped 230 control objectives into a self-assessment tool3, and the State Department integrates AI RMF concepts into its AI and Human Rights Profile.5
Many agencies lack dedicated AI compliance staff, so lightweight, cross-functional approaches are essential. Start with these four actions:
Policy emphasis may be shifting toward speed, but civil-rights considerations remain non-negotiable. Public administrators must ensure that the push for rapid adoption does not shortcut fairness analyses. The AI RMF's voluntary nature does not exempt agencies from Title VI, the Equal Credit Opportunity Act, or other legal frameworks.2 A cross-functional team, pulling in lawyers, data scientists, and program staff, can embed equity reviews into lightweight governance routines without stalling innovation. public service reforms offer instructive precedents for integrating accountability mechanisms into fast-moving institutional change.
These steps, though modest in resource demand, align with what is emerging as the federal standard and position agencies to adapt as mandates solidify. The proposed Federal AI Risk Management Act, though still in committee, signals that more formal requirements may be on the horizon.6
The federal government faces a stark tradeoff: invest heavily in AI talent to execute ambitious policy goals, or risk falling behind in a technology race where private-sector salaries and agility are formidable competitors. As AI spending and contracting surge, agencies can no longer afford to treat technical expertise as a niche specialty; it is a core requirement for responsible governance.
The Chief AI Officer (CAIO) is the linchpin of an agency's AI agenda. This role blends technology strategy, procurement oversight, risk management, and mission alignment. Under the Biden administration, the CAIO Council emphasized interagency coordination and careful risk assessment. The Trump-era council, with a smaller membership, prioritizes rapid adoption and competitive advantage. The composition of these councils matters: political appointees bring policy alignment but may lack technical depth, while career staff offer continuity but might be less responsive to shifting political directives. Liability remains a pressing, unresolved question. If an AI system causes harm or bias, does accountability rest with the CAIO, the agency head, or the vendor? Public administrators must clarify these lines to avoid paralyzing risk aversion.
Agency leaders do not need to code, but they do need a new suite of competencies. Core skills include:
Forward-looking public administration and policy schools are integrating AI governance into their curricula. Courses on algorithmic fairness, digital government, and tech procurement are moving from electives to core modules. For students and mid-career professionals, expertise in AI policy is rapidly becoming a powerful career differentiator in public administration. Programs accredited by NASPAA and others are beginning to respond, but the pace of change varies. Those who gain even foundational AI literacy in their degree programs will stand out in a job market hungry for technocrats who understand public values.
The federal government competes directly with the private sector for a limited pool of AI-literate talent. Salaries for experienced AI professionals can exceed $200,000, far above typical GS pay scales. To bridge the gap, agencies must invest in upskilling their existing workforce through university-local government MPA partnerships, paid training sabbaticals, and certification programs. The rapid expansion of AI contracts, from $4.6 billion in potential value in 2024 to $91.8 billion in 2026, underscores the urgency.1 Building internal capacity is not just a nice-to-have; it is a precondition for managing billions in taxpayer-funded AI initiatives responsibly.
With federal AI policy evolving rapidly from governance frameworks to execution-focused strategies, public administrators face pressing questions about implementation, spending, and workforce readiness. Drawing on a June 2026 Brookings analysis, here are answers to the most common inquiries shaping today's AI landscape in government.
Global AI spending is projected to reach $3.3 trillion in 2027, signaling that the scaling pressure on federal agencies will only intensify. For public administration careers and policymakers, this demands concrete preparation: stay current on Chief AI Officers Council directives, build internal capacity for AI auditing and risk assessment, and monitor federal-state preemption developments that could reshape compliance obligations. The rapid growth in contracts and the shift from risk-first to execution-first governance make one thing clear: adaptive governance structures that balance innovation with accountability are not optional. They are the defining professional challenge for this generation of public administrators.
