This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Official name: Promoting Advanced Artificial Intelligence Innovation and Security
Directs the Committee on National Security Systems, Secretary of War, and Homeland Security to prioritize AI-enabled cyber defense. Establishes AI cybersecurity clearinghouse. Develops classified benchmarking for AI models. Enhances AI capacity to protect against illegal computer access and exploitation.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is an Executive Order issued by the President of the United States with binding legal authority over federal agencies. It contains mandatory directives with specific timelines and enforcement mechanisms.
The document has good coverage of approximately 5-6 subdomains, with strong focus on AI system security (2.2), malicious actors using AI for cyberattacks (4.2), fraud and criminal manipulation (4.3), competitive dynamics (6.4), and dangerous capabilities (7.2). Coverage is concentrated in security, misuse prevention, and AI safety domains related to cybersecurity threats.
This Executive Order primarily governs AI use in Public Administration (excluding National Security) and National Security sectors, with extensive coverage of federal government cybersecurity systems. It also has minimal coverage of Finance and Insurance, Health Care, and Trade/Transportation/Utilities sectors through references to critical infrastructure operators like community banks, rural hospitals, and local utilities.
The document primarily focuses on the Deploy and Operate and Monitor stages of the AI lifecycle, with emphasis on secure deployment of frontier models and ongoing cybersecurity monitoring. It also addresses aspects of Verify and Validate through classified benchmarking processes for AI model capabilities.
The document explicitly mentions AI models, AI systems, and frontier models (covered frontier models). It establishes a classified benchmarking process to assess advanced cyber capabilities of AI models and sets thresholds for frontier model designation. The document focuses on AI-enabled cybersecurity capabilities and does not explicitly mention compute thresholds, open-weight models, general purpose AI, task-specific AI, generative AI, or predictive AI.
The document is an Executive Order issued by the President, as stated in the opening line and throughout the document's authority structure.
Multiple federal agencies and officials are designated with enforcement authority, including the Attorney General for criminal law enforcement and various agency heads for implementing cybersecurity directives and operational requirements.
The document establishes monitoring mechanisms through CISA for civilian systems, NSA for classified benchmarking of AI capabilities, and creates an AI cybersecurity clearinghouse to coordinate vulnerability scanning and remediation.
The document targets multiple federal agencies with mandatory cybersecurity requirements, and establishes voluntary collaboration frameworks with AI developers and critical infrastructure operators. The AI industry is explicitly mentioned as a target for voluntary engagement regarding frontier models.
5 subdomains (3 Good, 2 Minimal)
