NSA gets sweeping power to oversee private AI—with no public oversight
H.R. 3919 — Advanced AI Security Readiness Act · Filed by Darin LaHood (R-IL) · 5 cosponsors · Introduced Jun 11, 2025 · Referred to committee
Your members of Congress
Enter a ZIP to see where your representative and both senators stood on this bill.
Looked up on this device — your ZIP is never stored on our servers.
What it does
This bill directs the National Security Agency to develop an 'AI Security Playbook'—a classified and unclassified strategy document—to protect advanced AI systems from theft by foreign governments and well-resourced actors. The NSA must identify vulnerabilities in AI data centers, determine what AI components are most critical to protect, propose detection and prevention tactics, and analyze whether the U.S. government should directly build and oversee certain AI systems in secure government facilities. The bill requires the NSA to consult with private AI companies and report back to Congress within 90 and 270 days.
Why we flagged it
The bill establishes a formal NSA-led process to assess, secure, and potentially oversee advanced AI systems deemed critical to national security. While framed as defensive (protecting against theft), it creates a governance structure that may expand government involvement in private AI development and sets conditions for direct government control of certain AI systems.
What the text implies
- The bill authorizes the NSA to determine unilaterally which AI systems pose 'grave national security threats'—a definition that could expand to include commercial AI with no direct military application, giving the agency broad discretion to flag private companies' systems for government oversight.
- The 'hypothetical initiative to build covered AI technology systems in a highly secure governmental environment' language suggests a pathway toward government-owned AI development, potentially competing with or displacing private sector innovation and creating a dual-track AI ecosystem.
The full analysis lists 5 implications of this text.
Who stands to gain
Large AI developers (OpenAI, Anthropic, Google DeepMind, Meta AI Research); Cybersecurity and AI security consulting firms; Defense contractors with AI capabilities (Palantir, Booz Allen Hamilton, Raytheon Technologies)