On the heels of an executive order surrounding the security of AI systems, the White House released a Presidential Memorandum (NSPM-11) on Artificial Intelligence in the National Security Enterprise last Friday. The memorandum rescinds and replaces NSM-25 from the Biden Administration.

The purpose of the recent memorandum is to serve as a “new framework to put the most advanced, secure, and reliable AI systems into the hands of America’s warfighters and intelligence professionals while ensuring their responsible use,” according to the associated fact sheet.

Adding to the web of AI policy and guidance from the Trump Administration and Congress, the memorandum intersects with elements of the AI Action Plan and DOD DIB Roadmap to amplify innovative and advanced AI technologies in the national security sector through updated policy, key industry partnerships, buildout of enabling infrastructure, rapid acquisition, and expanding investment. It also outlines several next steps that are essential for contractors to monitor as AI systems continue to evolve within the national security landscape.

The order seeks to prioritize speed, sovereignty and mission effectiveness of AI among four “A” pillars:

• Adoption: Maintain rooted relationships with industry to identify mission areas where available AI frontier models can enhance operational effectiveness, while ensuring barriers to rapid deployment are removed.
• Adaption: Speedily adopt the most advanced commercial and open-source AI technologies across multiple vendors. Adopt deployment of commercially or internally customized AI technology in instances where commercial solutions are not appropriate due to security limitations.
• Assurance: Mandate that all adopted AI systems remain controllable and accountable, prohibiting any external entity from modifying or degrading systems that warfighters depend on. Implement rigorous testing, evaluation, validation and verification to ensure confidentiality, availability, reliability and interoperability of systems.
• Accountability: Ensure AI technologies are neither developed nor used outside the constitutional chain of command.

Next Steps and Milestones:

• Update Existing Autonomy Policy: Within 90 days, the Secretary of War will review and update DOD Directive 3000.09 on Autonomy in Weapon Systems to match evolving AI system capabilities.
• Noncompliant Contractor and Subcontractor Termination: The Secretary of War and Director of National Intelligence (DNI) are directed to terminate for default or for convenience those contracts which repeatedly demonstrate noncompliance with the memorandum’s requirements. This directive includes termination of subcontractors demonstrating noncompliance.
• Upcoming AI Governance Policy: Within 90 days, the Committee on National Security Systems and OMB shall issue policy, with a classified annex, on the governance of AI use in national security systems, including implementation and reporting requirements.
• Refresh Procurement Processes: Within 120 days, the Secretary of War, DNI and Intelligence Community (IC) heads will review and update procurement processes to ensure rapid onboarding of advanced AI models from multiple vendors.
• Infrastructure Access Framework: Within 90 days, Assistant to the President for Science and Technology (APST) and OMB to jointly develop a roadmap to access the most advanced computing resources. The framework should include commissioning advanced AI computing facilities, operating next-gen AI systems at scale, and establishing an AI test range.
• Partnerships Ensuring AI Security: Within 120 days, the Secretary of War, Secretary of Energy, DNI, and NSA Director should establish private-sector partnerships to secure AI technologies from cyber attacks, including conducting red-team exercises, sharing threat intelligence, and assisting with personnel vetting.
• Genesis Mission: Utilizing the Genesis Mission, the Secretary of Energy shall develop approaches to apply AI to national security missions.
• Foreign Technologies: The IC shall prioritize the analysis of foreign AI technologies across AI technology stacks, applications, use cases, governance and policies that pose a threat to the U.S.
• AI Data and Model Exchanges: Within 120 days, the Secretary of War, Secretary of Energy, DNI, and NSA Director will initiate joint AI data and model exchanges, to identify common AI applications across the national security sphere.
• Technical AI Talent: Within 120 days, OPM will initiate efforts to establish an AI National Security Strategic Reserve of non-governmental AI talent to provide federal support on AI national security issues, as needed.
• AI National Security Curriculum: Within 120 days, DNI and Secretary of War to develop an AI for National Security Curriculum to effectively train national security personnel to properly and securely employ AI systems.
• AI R&D: Agencies are directed to prioritize research and development of technologies that strengthen AI dependability, resilience, adaptability, and oversight.
• AI Risk Management Strategy: Within 120 days, DNI, Secretary of War and NSA Director will develop a joint strategy for AI risk management and establish AI security practices and standards for the national security enterprise.
• AI TEVV: Within 120 days, the Secretary of War will submit standardized AI national security Test, Evaluation, Verification, and Validation (TEVV) methodologies for review.

Implications for Contractors:

• Deepening Private-Sector Partnerships: The memorandum leaves little doubt of the private sector’s involvement to establish advanced AI technologies among national security agencies whilst weighing in on the barriers to adoption faced by vendors to procure the technology to the government.
• Increased and Varied AI Acquisition: The memorandum signals a push for new procurments of advanced AI models across multiple providers, including both large and small businesses.
• Prominence of AI System Integrity: The memorandum sets rigorous expectations for AI system dependability and control. Vendors supplying AI solutions to the government must provide offerings that withstand rigorous testing, evaluation, verification and validation.
• AI Infrastructure and Computing Facilities: The memorandum directs the buildout of next-generation, high-security computing facilities to run future AI systems at scale, representing potential new opportunities for contractors with capabilities related to data center construction and infrastructure, cloud architecture, and high-performance computing.
• AI R&D Priorities: The memorandum provides a clear signal to prioritize research and development investments on strengthening AI security and dependability; contractors should align AI-related R&D investments and offerings around those specific priorities.