Federal agencies across the United States are entering a new era of artificial intelligence (AI) adoption, thanks to two pivotal memos released by the White House’s Office of Management and Budget (OMB) on April 3, 2025. These revised policies—M-25-21 and M-25-22—are designed to accelerate responsible AI use and procurement, supporting the broader vision of Executive Order 14179 to remove barriers and boost American leadership in AI.
A Forward-Thinking Approach to AI in Government
Imagine a government where AI isn’t bogged down by red tape, but instead, is harnessed to deliver better public services, improve efficiency, and foster innovation. That’s the future the OMB envisions. The new memos replace earlier guidance, shifting from cautious experimentation to a “forward-leaning and pro-innovation” stance. The goal? To empower agencies, maximize U.S. competitiveness, and ensure the public benefits from AI advancements.
Key Shifts in Policy
- Streamlined Adoption: Agencies are now encouraged to cut unnecessary bureaucracy, making it easier to implement AI solutions quickly and effectively.
- Leadership and Governance: Each agency must appoint a Chief AI Officer (CAIO) to oversee AI strategy, risk management, and interagency collaboration. Governance boards will ensure cross-functional oversight, bringing together experts from IT, cybersecurity, data, and finance.
- Transparency and Public Trust: Agencies are required to publish annual inventories of their AI use cases and disclose risk assessments and waivers for high-impact AI applications. This transparency is designed to build public trust and demonstrate responsible AI use.
- Preference for American-Made Solutions: There’s a strong push to invest in U.S.-developed AI tools and nurture domestic AI talent, reinforcing America’s competitive edge.
What Does This Mean for Federal Agencies?
Accelerating Innovation
The OMB’s M-25-21 memo focuses on breaking down barriers to AI adoption. Agencies must develop strategies to remove obstacles, designate CAIOs within 60 days, and establish governance boards within 90 days. These steps are intended to foster a culture of innovation and ensure that AI is used responsibly and effectively.
Managing Risks and Ensuring Oversight
Not all AI is created equal. For “high-impact” AI—systems that significantly affect rights, safety, or legal outcomes—agencies must conduct thorough risk assessments, pre-deployment testing, and ongoing monitoring. Human oversight is mandatory, and individuals affected by AI decisions must have access to remedies or appeals. If an AI system doesn’t meet minimum standards, its use must be paused or stopped.
Building a Skilled Workforce
Agencies are encouraged to invest in AI training and upskilling, ensuring their teams are ready to manage and govern AI technologies. Clear expectations and accountability channels are set to help staff navigate the evolving landscape of AI use.
Smarter, More Responsible AI Procurement
The companion memo, M-25-22, addresses how agencies should acquire AI systems. The focus is on:
- Fostering a Competitive Marketplace: Agencies are urged to buy American-made AI and develop in-house expertise.
- Protecting Privacy and Intellectual Property: Contracts must safeguard government data and prevent unauthorized use for AI training or development.
- Ensuring Value and Performance: Procurement processes should encourage competition, prevent vendor lock-in, and require regular performance assessments.
- Knowledge Sharing: Agencies will contribute to a shared repository of best practices, helping the entire government learn and improve together.
Actionable Takeaways for Agencies and Stakeholders
- Designate a CAIO and establish governance boards promptly.
- Develop clear, public-facing AI use inventories and risk reports.
- Invest in workforce training and upskilling for AI readiness.
- Prioritize American-made AI solutions and talent development.
- Implement robust risk management and oversight for high-impact AI.
Frequently Asked Questions
Q: What are the main goals of the new OMB AI policies?
A: The new OMB policies aim to accelerate responsible AI adoption in federal agencies by reducing bureaucratic barriers, enhancing governance, ensuring transparency, and prioritizing American-made AI solutions.
Q: How do the memos address AI risk management?
A: Agencies must implement risk management practices for high-impact AI use cases, including pre-deployment testing, impact assessments, human oversight, and remedies for affected individuals.
Q: What is the role of Chief AI Officers (CAIOs) in federal agencies?
A: CAIOs are responsible for leading AI governance, risk management, and strategic adoption efforts within their agencies, as well as coordinating with other agencies and stakeholders.
Q: How do the new policies promote transparency in government AI use?
A: Agencies are required to annually publish inventories of their AI use cases and publicly report risk determinations and waivers for high-impact AI, including justifications.
Q: How do the memos support American AI innovation and talent?
A: The memos encourage agencies to invest in U.S.-developed AI solutions and develop or retain American AI talent, fostering a competitive and innovative domestic AI marketplace.
In Summary
- The OMB’s new memos mark a significant shift toward rapid, responsible AI adoption in government.
- Agencies must streamline processes, appoint AI leadership, and ensure robust governance.
- Transparency and public trust are central, with annual reporting and risk disclosures.
- American-made AI solutions and talent are prioritized to boost national competitiveness.
- Risk management and oversight are mandatory for high-impact AI, protecting both agencies and the public.