White House AI Action Plan: Progress and What Executives Need to Do Now
Since July, the federal government has moved from AI talking points to execution. Agencies are rewriting procurement playbooks, standing up governance, and taking a harder line on state rules that could interfere with national AI policy.
If you lead strategy or budget, here's what changed, why it matters, and the near-term actions to keep your org compliant and competitive.
The Plan in 30 Seconds
- 90+ federal actions across three pillars: accelerate AI innovation, build American AI infrastructure, and lead in international AI diplomacy and security.
- Follow-on moves hit procurement, regulation, standards, and agency-level AI strategies.
- Outcome to track: faster AI deployment with stronger oversight and clearer guardrails.
Federal Pushback on Restrictive State AI Laws
In December, an executive order directed an AI Litigation Task Force to review and potentially challenge state AI laws deemed "unconstitutional, preempted, or otherwise unlawful." The aim is to remove friction that could stall national AI objectives and private-sector innovation.
What this means for enterprises: compliance strategies may centralize around federal standards. Expect more preemption arguments and less state-by-state fragmentation over time, especially for high-impact models and public-sector contracts.
Procurement Rules You'll Feel This Quarter
OMB released procurement guidance tied to the AI Action Plan with specific LLM expectations. The memo pushes for unbiased AI principles, including truth-seeking and ideological neutrality, and clarifies what agencies should and shouldn't request from vendors. Procurement teams should review the AI Learning Path for Procurement Specialists to align templates and vendor requirements with the new guidance.
- Required in LLM solicitations: acceptable use policy; model, system, or data cards; end-user resources; and mechanisms for end-user feedback.
- What to avoid: compelling disclosure of model weights and other sensitive technical details.
- Deadline: agencies must update procedures and policies by March 11.
Leader actions now:
- Update RFP templates and vendor questionnaires to align with the new OMB requirements.
- Stand up a review lane for acceptable use policies, model/documentation cards, and feedback workflows.
- Codify a stance on protected IP and model-weight nondisclosure in your contracting standards.
- Establish a process to validate "truth-seeking" and ideological neutrality claims through audits or third-party assessments.
Agency Strategies to Watch
Agencies are aligning with the plan while tailoring execution to their missions. These documents are not shelfware; they are driving budgets, pilots, and oversight.
- HHS: Governance, workforce readiness, risk management, and modernization to improve health and human services.
- VA: AI to streamline workflows, improve care delivery, accelerate benefits processing, and build public trust. Focus areas include data infrastructure, workflow redesign, and an AI-ready workforce.
- State Department: Enterprise Data and AI Strategy for 2026, with two goals: pioneering advanced statecraft and accelerating adoption across operations.
Army CIO: Treat Prompts as Records
Guidance issued in August requires AI products used in the Army to comply with records management, privacy rules, and FOIA. Prompts are treated as records and must be captured with context for compliant retention and discovery.
- Capture prompts, outputs, versions, datasets referenced, and decision rationale where applicable.
- Map retention schedules and access controls to protect sensitive data while enabling auditability.
- Train users on what constitutes a record during AI-assisted work.
What This Means for Your Roadmap (Next 60-90 Days)
- Refactor procurement artifacts to meet OMB's LLM requirements before March 11.
- Run a gap assessment on AI governance: acceptable use, feedback loops, incident reporting, and model documentation standards.
- Stand up a data and records policy for prompts and outputs, including FOIA and privacy controls.
- Prioritize pilots that deliver measurable efficiency gains in benefits processing, case management, and knowledge workflows.
- Align vendor contracts to protect IP while enabling transparency for evaluations and audits.
- Upskill leadership and critical functions on AI risk, policy, and procurement changes.
Risks and Opportunities
- Compliance risk: Missing the March 11 updates will slow awards and increase review cycles.
- Model risk: Claims of neutrality and truth-seeking require evidence. Build evaluation and monitoring plans you can defend.
- Operational upside: Agencies are signaling demand for AI that shortens time-to-service and improves decision quality. Vendors that meet the documentation and governance bar will move faster.
If You Lead Strategy, Here's the Short List
- Finalize a unified AI procurement playbook with model-card and acceptable-use requirements.
- Implement prompt and output logging where AI influences decisions or public communications.
- Direct a cross-functional review (legal, security, privacy, operations) to validate neutrality and truth-seeking claims.
- Pick two high-ROI pilots and instrument them with clear metrics for cost, time, and quality.
- Brief the board or executive team on the federal preemption push and its impact on state compliance plans; for executive oversight and governance training, see the AI Learning Path for CIOs (Chief Information Officers).
Helpful Resources
The signal is clear: procurement is tightening, governance is getting specific, and state-by-state friction may ease as federal standards take center stage. Move early, document well, and build evaluation into your operating rhythm.
Your membership also unlocks:
