Defines AI systems and outlines U.S. regulations for investments in national security technologies with countries of concern. Requires notification of certain AI-related transactions, classifies them as notifiable or prohibited and mandates compliance oversight, reporting, and potential penalties for violations.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding federal regulation issued by the Department of the Treasury under the International Emergency Economic Powers Act (IEEPA) with explicit enforcement mechanisms, civil and criminal penalties, and mandatory compliance requirements for U.S. persons.
The document has good coverage of approximately 6-8 subdomains, with strong focus on malicious actors (4.1, 4.2), AI system security (2.2), competitive dynamics (6.4), governance failure (6.5), and AI safety failures (7.1, 7.2, 7.3). Coverage is concentrated in security, misuse prevention, national security, and AI safety domains.
This regulation primarily governs the Information sector (AI development, semiconductors, data processing) and Scientific Research and Development Services sector (quantum computing, advanced AI research). It also has significant coverage of National Security sector activities. The regulation applies to U.S. persons making investments in these technologies in countries of concern.
The document covers multiple AI lifecycle stages with primary focus on Build and Use Model (AI system development with compute thresholds), Deploy (investment transactions), and Operate and Monitor (post-transaction knowledge requirements). It addresses planning through notification requirements and verification through due diligence standards.
The document explicitly defines and regulates AI systems with detailed technical specifications. It establishes compute thresholds at 10^23, 10^24, and 10^25 FLOPs for different categories. It covers AI systems broadly but does not explicitly distinguish between general purpose AI, task-specific AI, foundation models, or generative vs predictive AI. It does not explicitly address open-weight models.
Department of the Treasury, Secretary of the Treasury
The regulation was issued by the Department of the Treasury implementing Executive Order 14105. The Secretary of the Treasury is explicitly identified as the authority who issues these regulations in consultation with other agencies.
Department of the Treasury, Secretary of the Treasury, Assistant Secretary of the Treasury for Investment Security, Attorney General, Department of Justice
The Department of the Treasury has primary enforcement authority including investigation, penalty assessment, and divestment orders. The Attorney General and Department of Justice handle criminal prosecutions and certain civil enforcement actions.
Department of the Treasury, Secretary of Commerce, Secretary of State, Secretary of Defense, Secretary of Energy, Director of National Intelligence, relevant agencies
The Department of the Treasury monitors compliance through notification requirements, investigations, and information requests. The Secretary consults with multiple agencies on analysis of transactions and national security implications.
U.S. persons (including U.S. citizens, lawful permanent residents, entities organized under U.S. laws, and persons in the United States), controlled foreign entities of U.S. persons, covered foreign persons engaging in covered activities in countries of concern
The regulation applies to U.S. persons who engage in transactions with covered foreign persons. It regulates investments in AI systems, semiconductors, and quantum technologies. U.S. persons include entities that develop, deploy, or provide infrastructure for AI.
7 subdomains (4 Good, 3 Minimal)