Executive Order 14110 on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence

Executive Order 14110 establishes comprehensive federal policy for the safe, secure, and trustworthy development and use of artificial intelligence across the United States government and economy. The order articulates eight guiding principles for AI governance, including requirements that AI must be safe and secure, promote responsible innovation and competition, support American workers, advance equity and civil rights, protect consumers, safeguard privacy, strengthen federal AI capabilities, and enhance American leadership globally.

The order creates extensive new requirements for AI developers, particularly those working with dual-use foundation models that exceed specified computational thresholds. Companies must report to the department-of-commerce on training activities, model weights security, and ai-red-teaming results. The order establishes new reporting requirements for large-scale computing clusters and mandates safety evaluations before deployment.

Key institutional changes include the creation of the white-house-artificial-intelligence-council to coordinate federal AI policy, requirements for agencies to designate Chief AI Officers, and the establishment of AI governance boards. The order directs nist to develop new guidelines for AI safety and security, including companion resources to the nist-ai-rmf-1-0 for generative-ai.

The order addresses AI risks across critical sectors including healthcare, transportation, education, and critical infrastructure. It establishes new frameworks for managing AI risks in the criminal justice system, protecting civil rights, and ensuring consumer protection. The order also creates new pathways for attracting AI talent to government and establishes requirements for federal agencies' use of AI systems.

International cooperation features prominently, with directives to strengthen global AI governance frameworks, develop international standards, and promote responsible AI development abroad. The order builds on existing frameworks including the nist-ai-rmf-1-0, the Blueprint for an AI Bill of Rights, and various sector-specific guidance documents.