Instructs the NSA Director to develop AI security guidance focusing on cybersecurity risks and unique challenges. Requires collaboration with AI experts and other departments, including roundtables and expert panels. Mandates initial and final progress reports to congressional committees.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding legislative act from the United States Congress with mandatory obligations on the NSA Director, specified timelines, and reporting requirements to congressional committees.
The document has good coverage of approximately 6-8 subdomains, with strong focus on AI system security (2.2), malicious actors (4.1, 4.2), competitive dynamics (6.4), governance failure (6.5), and AI safety failures (7.2, 7.3). Coverage is concentrated in security, threat prevention, and system robustness domains.
This document primarily governs the National Security sector through mandates on the NSA. It also has significant implications for the Information sector (AI developers), Scientific Research and Development Services (AI researchers and National Laboratories), and Professional and Technical Services (consultants and experts engaged in the process). The guidance developed will ultimately apply to any sector developing or deploying advanced AI with national security implications.
The document primarily focuses on the Deploy and Operate and Monitor stages, with emphasis on securing AI systems post-development. It addresses security measures throughout the AI supply chain including model weights protection, but does not extensively cover early-stage planning, data collection, or initial model building phases.
The document explicitly mentions AI systems, AI models, and AI technologies throughout. It focuses on advanced AI with critical capabilities and covers the entire AI supply chain including model weights, training data, and computing environments. It does not explicitly mention frontier AI, general purpose AI, foundation models, generative AI, or specific compute thresholds.
United States Congress; Senate; House of Representatives
The document is a Congressional bill proposed by the United States Congress, as indicated by the opening text 'Be it enacted by the Senate and House of Representatives of the United States of America in Congress assembled'.
Congressional intelligence committees; Select Committee on Intelligence of the Senate; Permanent Select Committee on Intelligence of the House of Representatives
The congressional intelligence committees serve as the enforcement mechanism through mandatory reporting requirements and oversight. The NSA Director must submit reports to these committees at specified intervals.
Congressional intelligence committees; Select Committee on Intelligence of the Senate; Permanent Select Committee on Intelligence of the House of Representatives
The congressional intelligence committees monitor implementation through required progress reports at 180 days and final reports at 365 days, including summaries of progress and remaining work.
Director of the National Security Agency; Artificial Intelligence Security Center; prominent artificial intelligence developers and researchers; service providers; private sector
The primary target is the NSA Director who must develop guidance. Secondary targets include AI developers and researchers who will be engaged in the process and service providers who will receive security briefings. The guidance is also intended for dissemination to the private sector.
9 subdomains (5 Good, 4 Minimal)