United States Executive Orders on Artificial Intelligence

Executive orders on artificial intelligence are directives issued by the President of the United States that instruct federal agencies on how to develop, procure, regulate and deploy AI. Because the US has no comprehensive federal AI statute, executive orders — together with agency rulemaking — have functioned as the country's main lever of national AI policy, and the direction of that lever has reversed sharply between administrations.^[1][2]

History

The modern wave began with President Biden's Executive Order 14110 of 30 October 2023, Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence — the most sweeping US AI order to date. It required developers of the most powerful models to share safety test results with the government, invoked the Defense Production Act over frontier training runs, directed standards work on watermarking and synthetic content, and touched on privacy, civil rights, healthcare and workforce impacts.^[1]

Within days of taking office in January 2025, President Trump revoked EO 14110 through Executive Order 14148, and on 23 January 2025 signed Executive Order 14179, Removing Barriers to American Leadership in Artificial Intelligence, which ordered a review and removal of policies enacted under the revoked order on the grounds that they acted as barriers to innovation.^[2] In July 2025 the administration released America's AI Action Plan, a 25-page strategy built around deregulation, export promotion of US technology stacks and "ideologically neutral" AI, accompanied by three further executive orders — including measures on federal procurement and on accelerating data-centre and energy permitting.^[3][4]

Key Concepts

• Speed over statute. Executive orders take effect immediately but bind only the executive branch; they can be (and have been) revoked wholesale by the next president, making them inherently volatile policy.^[2]
• Procurement as regulation. Much of the 2025 approach works through the government's purchasing power: what federal agencies are allowed to buy, from whom, and under which conditions.^[3]
• Pre-emption concerns. Federal efforts in 2025–2026 have repeatedly targeted state AI laws, seeking to prevent a patchwork of divergent state regulations — a contrast with the EU, which legislated comprehensively through the AI Act.^[4]
• Export and infrastructure dimensions. Modern orders couple model policy with chip export controls, data-centre permitting and grid capacity, treating compute as strategic infrastructure.^[3]

Applications and Impact

Executive orders shape not only federal use of AI but private-sector practice, because agencies translate them into procurement rules, safety-institute mandates, NIST standards work and funding conditions. The 2023 order pushed the AI Safety Institute (later renamed the Center for AI Standards and Innovation) and safety-testing commitments for frontier models; the 2025 turn loosened those requirements while emphasising deployment speed in defence, energy and industry.^[1][3] For international observers, the US sequence demonstrates both the responsiveness and the fragility of executive-led AI governance.

>See Also

• EU AI Act
• AI Regulation in Malaysia
• Malaysia AI Governance Framework
• National AI Office Malaysia
• AI Safety