Directs the federal government to ensure the safety and security of AI systems and take action to protect Americans from their potential risks. Prescribes measures to improve American innovation, competition, and leadership in AI while preserving privacy, mitigating algorithmic discrimination, and addressing labor disruptions. Calls for further responsible deployment of AI in federal government and consumer-facing industries.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding Executive Order issued by the President with legal authority under the Constitution, the Defense Production Act, and the International Emergency Economic Powers Act. It contains mandatory language throughout, establishes enforcement mechanisms, and creates binding obligations on federal agencies.
The document has good coverage of approximately 15-18 subdomains, with strong focus on AI system security (2.2), malicious actors (4.1, 4.2, 4.3), privacy compromise (2.1), misinformation (3.1, 3.2), discrimination (1.1, 1.3), governance failure (6.5), competitive dynamics (6.4), and AI safety failures (7.1, 7.2, 7.3, 7.4). Coverage is concentrated in security, misuse prevention, civil rights, and AI safety domains.
This Executive Order governs AI use across virtually all sectors of the U.S. economy, with particularly detailed coverage of Information (AI development companies), National Security, Public Administration, Health Care and Social Assistance, Finance and Insurance, and Educational Services. It also addresses critical infrastructure sectors including utilities, transportation, and telecommunications.
The document comprehensively addresses all stages of the AI lifecycle, with particularly strong emphasis on Build and Use Model (through development requirements and red-teaming), Verify and Validate (through testing and evaluation requirements), Deploy (through reporting and safety requirements), and Operate and Monitor (through ongoing monitoring and incident reporting). It also addresses Plan and Design through strategic planning requirements and Collect and Process Data through data quality and privacy provisions.
The document explicitly addresses AI models and AI systems with detailed definitions. It extensively covers dual-use foundation models with specific compute thresholds (10^26 FLOPs). It addresses generative AI throughout. It mentions open model weights in the context of dual-use foundation models. The document does not explicitly use the terms 'frontier AI,' 'general purpose AI,' 'task-specific AI,' 'foundation models' (though dual-use foundation models are covered), or 'predictive AI.'
Executive Office of the President, President Joseph R. Biden Jr.
The Executive Order is issued by the President of the United States under constitutional and statutory authority, making the Executive Office of the President the proposer of this governance instrument.
Department of Commerce, Department of Justice, Department of Homeland Security, Department of Health and Human Services, Federal Trade Commission, Office of Management and Budget, independent regulatory agencies, Sector Risk Management Agencies
Multiple federal agencies are designated with enforcement authority, including regulatory agencies with specific mandates. The Secretary of Commerce has explicit authority under the Defense Production Act and International Emergency Economic Powers Act.
Office of Management and Budget, White House Artificial Intelligence Council, Department of Homeland Security (Artificial Intelligence Safety and Security Board), National Institute of Standards and Technology, Chief Data Officer Council, agency Chief AI Officers, interagency working groups
The order establishes multiple monitoring mechanisms including a White House AI Council, agency-level Chief AI Officers, and various reporting requirements to OMB and other coordinating bodies.
Companies developing dual-use foundation models, United States Infrastructure as a Service Providers, Federal agencies, critical infrastructure owners and operators, healthcare providers, financial institutions, employers, State and local governments
The order targets multiple entity types: private sector AI developers (especially those creating dual-use foundation models), federal agencies using AI, and various deployers across sectors including healthcare, finance, transportation, and critical infrastructure.
21 subdomains (15 Good, 6 Minimal)