White House Publishes National Security Presidential Memorandum, Removing Guardrails on AI

On June 5, 2026, the Trump administration issued a National Security Presidential Memorandum (NSPM 11) on Artificial Intelligence in the National Security Enterprise. This memorandum furthers the Trump administration’s efforts to accelerate the adoption and deployment of AI systems in national security by instructing the heads of multiple agencies working on national security, military operations, intelligence, and government data processing to adapt policies and contracts to that effect. 

While NSPM 11 pays lip service to ensuring that the deployment of AI systems is “consistent with” constitutional rights, civil liberties, and laws and regulations protecting the privacy of Americans, it provides little information to agencies on how specifically to accomplish such a goal—indeed, the majority of NSPM 11’s content runs contrary to it. This memo rescinds and replaces National Security Memorandum-25 (NSM 25), a Biden administration memo that emphasized the need to protect human rights, civil rights, civil liberties, and privacy when using AI. NSM 25 was published along with a framework detailing concrete steps for the federal government to take to protect rights and liberties while deploying AI for national security systems. The framework prohibited certain uses, identified high-impact uses that require additional scrutiny, and outlined the minimal risk management practices agencies must implement. But the new NSPM 11 provides no such framework to protect fundamental rights, despite the dangers posed by AI. 

Furthermore, NSPM 11’s reference to making AI deployments consistent with laws and regulations “governing the privacy and civil liberties of American citizens” raises serious concerns. To be clear, constitutional rights, civil liberties, and privacy protections apply to U.S. persons, including noncitizens. The executive branch has no authority to ignore the rights of U.S. persons in any context, including deploying AI systems in national security. 

EPIC has long been an advocate for stronger oversight and regulation of AI systems in government, including in national security.