United States - National AI Policy Framework

Overview

The Executive Order on Ensuring a National AI Policy Framework, issued in December 2025, represents a landmark directive from the President of the United States aimed at establishing a comprehensive and cohesive national strategy for the development, deployment, and governance of artificial intelligence. This order acknowledges the transformative potential of AI across all sectors of society, from economic growth and scientific discovery to national security and public services, while simultaneously recognizing the profound ethical, safety, and societal challenges it presents. The primary objective of this framework is to foster responsible innovation, protect American values, and maintain the nation's competitive edge in the global AI landscape. It mandates a whole-of-government approach, requiring federal agencies to collaborate and coordinate their efforts to achieve these overarching goals, ensuring that AI development is human-centric, trustworthy, and aligned with democratic principles. The order emphasizes the need for a balanced approach that encourages technological advancement while implementing robust safeguards against potential harms, including issues related to bias, privacy, security, and accountability. It sets the stage for future regulatory actions and policy developments, providing a foundational structure upon which more specific rules and guidelines will be built, thereby shaping the trajectory of AI in the United States for decades to come.

Furthermore, this Executive Order underscores the urgency of addressing AI-related challenges proactively, rather than reactively. It recognizes that the rapid pace of AI innovation necessitates a dynamic and adaptive policy response that can evolve alongside technological advancements. The framework outlines key pillars for action, including promoting responsible AI innovation, safeguarding American consumers and workers, advancing equity and civil rights, protecting privacy, ensuring national security, and fostering international cooperation. By establishing clear directives and responsibilities for various federal departments and agencies, the order seeks to streamline the government's approach to AI, eliminate redundancies, and ensure a consistent application of principles across different domains. It also calls for significant investment in AI research and development, workforce training, and public education to ensure that the United States remains at the forefront of AI innovation and that its citizens are equipped to thrive in an AI-driven future. The Executive Order serves as a critical step in translating broad policy aspirations into concrete governmental actions, laying the groundwork for a robust and resilient national AI ecosystem that benefits all Americans.

Definitions

For the purposes of this Executive Order, 'Artificial Intelligence' or 'AI' refers to a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. AI systems utilize machine and human-based inputs to perceive real and virtual environments; abstract such perceptions into models through analysis in an automated manner; and use model inference to formulate options for information or action. This broad definition encompasses various sub-fields such as machine learning, deep learning, natural language processing, computer vision, and robotics, recognizing the diverse applications and underlying technologies that fall under the AI umbrella. The order distinguishes between general-purpose AI systems and those designed for specific, narrow tasks, acknowledging that different levels of scrutiny and regulatory approaches may be necessary depending on the system's capabilities, intended use, and potential impact. The definition is intended to be technology-neutral and future-proof, allowing the framework to remain relevant as AI technologies continue to evolve and new capabilities emerge, ensuring that the policy framework can adapt to unforeseen advancements without requiring constant redefinition of its core subject matter.

The Executive Order also defines 'High-Risk AI Systems' as those AI systems that pose a significant risk of harm to the fundamental rights, safety, or well-being of individuals, or to critical infrastructure, national security, or democratic institutions. This classification is crucial as it triggers enhanced requirements for risk management, transparency, testing, and oversight. Examples of high-risk AI applications might include those used in critical infrastructure management, law enforcement, employment decisions, credit scoring, healthcare diagnostics, and autonomous vehicles. The determination of what constitutes 'high-risk' is based on a combination of factors, including the severity and probability of potential harm, the context of use, and the availability of human oversight and mitigation measures. The order mandates that federal agencies develop specific criteria and methodologies for identifying and assessing high-risk AI systems within their respective jurisdictions, ensuring a consistent and robust approach to risk classification across the government. Furthermore, 'Responsible AI Innovation' is defined as the development and deployment of AI systems in a manner that prioritizes safety, security, privacy, fairness, transparency, and accountability, while simultaneously fostering economic growth and technological advancement. This definition guides the ethical considerations and operational principles that underpin the entire national AI policy framework, emphasizing a balance between innovation and protection.

Governance and Institutional Framework

The Executive Order establishes a robust governance and institutional framework designed to ensure coordinated and effective implementation of the national AI policy. At its core, it mandates the creation of a National AI Advisory Council (NAIAC), composed of leading experts from academia, industry, civil society, and government, tasked with providing ongoing strategic advice to the President and federal agencies on AI policy, research priorities, and ethical considerations. This council will serve as a critical interface between the government and external stakeholders, ensuring that policy decisions are informed by a diverse range of perspectives and expertise. Furthermore, the order designates the Office of Science and Technology Policy (OSTP) as the lead coordinating body for interagency AI policy development, responsible for harmonizing efforts across federal departments and agencies, developing common standards, and reporting on progress. OSTP will work in close collaboration with the National Security Council (NSC) and the National Economic Council (NEC) to ensure that AI policy aligns with broader national security and economic objectives, reflecting the cross-cutting nature of AI's impact. This multi-layered governance structure is designed to provide both strategic direction and operational oversight, ensuring that the national AI policy framework is consistently applied and continuously refined.

To further bolster the federal government's capacity to manage and regulate AI, the Executive Order directs each federal agency with significant AI responsibilities to appoint a Chief AI Officer (CAIO) or designate an existing senior official to fulfill this role. These CAIOs will be responsible for overseeing their respective agency's AI strategy, ensuring compliance with the national framework, promoting responsible AI adoption within their operations, and serving as a point of contact for interagency coordination. This decentralized yet coordinated approach aims to embed AI expertise and accountability within each relevant agency, allowing for tailored implementation while adhering to overarching national principles. The order also mandates the establishment of an interagency AI working group, co-chaired by OSTP and the National Institute of Standards and Technology (NIST), to develop and update technical standards, benchmarks, and best practices for AI safety, security, and trustworthiness. This working group will play a crucial role in translating policy objectives into actionable technical guidance, facilitating consistent implementation across the government and providing a clear reference for industry. The comprehensive governance structure outlined in this Executive Order is intended to create a durable and adaptable system for managing the opportunities and risks presented by AI, ensuring long-term national leadership and responsible development.

Key Focus Areas

The Executive Order outlines several key focus areas that form the pillars of the national AI policy framework, each addressing a critical dimension of AI's societal impact. One paramount focus area is the promotion of responsible AI innovation. This involves fostering an environment that encourages groundbreaking research and development while embedding ethical considerations and safety measures from the design phase. The order directs federal agencies to prioritize funding for AI research that not only advances technological capabilities but also explores methods for mitigating bias, enhancing transparency, and ensuring robust security. It emphasizes the importance of public-private partnerships to accelerate innovation, facilitate knowledge transfer, and develop shared infrastructure for AI testing and evaluation. Furthermore, the framework aims to reduce barriers to AI adoption for small businesses and startups, ensuring that the benefits of AI innovation are broadly distributed across the American economy. This focus area is critical for maintaining the United States' competitive edge in AI, ensuring that innovation is not stifled by overly burdensome regulations but rather guided by principles of responsible development and deployment, leading to beneficial outcomes for society as a whole.

Another critical focus area is safeguarding American consumers and workers from the potential negative impacts of AI. This includes directives to protect against algorithmic discrimination, ensure fair and equitable access to opportunities, and address the implications of AI on the future of work. The order mandates that agencies develop guidelines and enforcement mechanisms to prevent AI systems from perpetrating or exacerbating existing societal biases, particularly in sensitive areas such as employment, housing, and credit. It also calls for proactive measures to support the American workforce, including investments in AI-related education and training programs, to equip workers with the skills needed for an evolving job market. This focus area recognizes that while AI offers immense benefits, it also poses risks to individual rights and economic stability, necessitating robust protections. Additionally, the Executive Order places significant emphasis on ensuring national security and protecting critical infrastructure from AI-related threats. This involves developing strategies to counter malicious AI use, enhancing cybersecurity defenses for AI systems, and leveraging AI to strengthen national defense capabilities responsibly. The framework seeks to strike a delicate balance between harnessing AI for strategic advantage and mitigating the risks associated with its misuse by adversarial actors, ensuring the nation's security in an increasingly AI-driven world.

Implementation Framework

The implementation framework articulated in the Executive Order is designed to translate the broad policy objectives into concrete, actionable steps across the federal government. A central component of this framework is the mandate for federal agencies to conduct comprehensive AI impact assessments for all new or significantly modified AI systems they develop, procure, or utilize. These assessments must identify potential risks related to privacy, security, bias, and societal impact, and outline mitigation strategies before deployment. This proactive approach ensures that potential harms are identified and addressed early in the AI lifecycle, rather than after systems are already in operation. Furthermore, the order directs the National Institute of Standards and Technology (NIST) to develop and publish a series of AI risk management frameworks and technical standards, building upon its existing work in this area. These frameworks will provide agencies and private sector entities with practical guidance on how to assess, manage, and communicate AI-related risks, fostering a common language and approach to AI safety and trustworthiness across diverse applications. The implementation framework emphasizes a continuous improvement model, where lessons learned from early deployments and impact assessments inform subsequent policy updates and technical guidance, ensuring the framework remains adaptive and effective.

Another key aspect of the implementation framework involves the establishment of AI testing and evaluation protocols. The Executive Order mandates that federal agencies, particularly those dealing with high-risk AI systems, develop and implement rigorous testing methodologies to ensure the reliability, robustness, and fairness of AI systems before and during deployment. This includes requirements for independent third-party evaluations where appropriate, to enhance public trust and verify compliance with established standards. The order also encourages the creation of AI 'sandboxes' or regulatory testbeds, which will allow for the safe experimentation and development of innovative AI technologies under controlled conditions, facilitating learning and adaptation without immediately imposing full regulatory burdens. These sandboxes are intended to accelerate the responsible development of cutting-edge AI solutions by providing a flexible environment for testing and iteration. Moreover, the implementation framework calls for enhanced transparency requirements, particularly for high-risk AI systems, including clear documentation of system design, data sources, performance metrics, and decision-making processes. This transparency is crucial for accountability, allowing for effective oversight and enabling individuals to understand how AI systems affect them. The comprehensive implementation framework outlined in the Executive Order aims to create a structured and accountable process for integrating AI into government operations and guiding its responsible development in the private sector.

Monitoring and Evaluation

The Executive Order establishes a robust system for the ongoing monitoring and evaluation of the national AI policy framework's effectiveness and impact. Central to this system is the requirement for the Office of Science and Technology Policy (OSTP) to submit an annual report to the President and Congress detailing the progress of federal agencies in implementing the directives of the Executive Order. This report will include an assessment of the overall state of AI development and deployment in the United States, identify emerging risks and opportunities, and recommend any necessary adjustments to the policy framework. The annual reporting mechanism ensures continuous oversight and provides a regular opportunity to review and update the national strategy in response to technological advancements and evolving societal needs. Furthermore, federal agencies are mandated to establish internal mechanisms for monitoring the performance and impact of AI systems within their purview, including tracking incidents of algorithmic bias, security breaches, or other adverse outcomes. This internal monitoring is critical for identifying potential problems early and implementing corrective actions, fostering a culture of continuous improvement and accountability within government operations. The emphasis on data-driven evaluation ensures that policy decisions are informed by empirical evidence and real-world outcomes.

In addition to internal agency monitoring, the Executive Order calls for the establishment of independent review processes for high-risk AI systems. This may involve the creation of external expert panels or the engagement of independent auditors to conduct periodic assessments of these systems, verifying their compliance with safety, fairness, and transparency standards. The goal of these independent reviews is to provide an additional layer of assurance and to build public trust in the responsible deployment of AI, particularly in sensitive sectors. The order also directs the National AI Advisory Council (NAIAC) to provide regular public assessments of the national AI policy framework, offering recommendations for improvements based on their expertise and engagement with stakeholders. This public accountability mechanism ensures transparency and allows for broader societal input into the ongoing evolution of AI policy. Moreover, the framework encourages academic institutions and civil society organizations to conduct independent research on the societal impacts of AI and to contribute to the ongoing evaluation efforts, fostering a collaborative ecosystem for understanding and addressing AI's complex challenges. Through this multi-faceted approach to monitoring and evaluation, the Executive Order aims to create a dynamic and responsive national AI policy that can adapt to the rapid pace of technological change and effectively address both the opportunities and risks presented by artificial intelligence.

Penalties, Liability, and Appeals

While the Executive Order primarily focuses on establishing a policy framework and directing federal agencies, it lays the groundwork for future regulatory actions concerning penalties, liability, and appeals related to AI systems. The order directs relevant agencies, such as the Department of Justice, the Federal Trade Commission, and sector-specific regulators, to review existing legal and regulatory frameworks to identify gaps and propose legislative or regulatory changes necessary to address AI-specific issues of liability and enforcement. This includes examining how current product liability laws, tort law principles, and consumer protection statutes apply to AI systems, particularly in cases where autonomous AI decisions lead to harm. The intent is to ensure that there are clear mechanisms for accountability when AI systems cause damage, whether due to design flaws, data biases, or misuse. The order emphasizes that the lack of clear liability rules should not impede the development of beneficial AI, but rather that a predictable legal environment is essential for fostering responsible innovation and protecting affected parties. It anticipates that future legislative actions will establish specific penalty regimes for non-compliance with AI safety and ethical standards, potentially including fines, operational restrictions, and other sanctions, particularly for high-risk AI systems that fail to meet mandated requirements. The framework recognizes that effective enforcement is crucial for ensuring adherence to the national AI policy and maintaining public trust.

Regarding appeals and redress, the Executive Order directs agencies to develop mechanisms that allow individuals to challenge decisions made or significantly influenced by AI systems, especially in contexts that affect their rights, access to services, or economic opportunities. This includes ensuring that individuals have the right to an explanation of AI-driven decisions, the ability to request human review, and access to effective remedies when an AI system's decision is found to be unfair, discriminatory, or erroneous. The order emphasizes the importance of due process and transparency in AI-assisted decision-making, particularly in government services and other public-facing applications. Agencies are tasked with identifying existing appeal processes that can be adapted for AI-related grievances and, where necessary, proposing new administrative or judicial avenues for redress. This focus on appeals and redress mechanisms is critical for upholding fundamental rights and ensuring that AI systems are used in a manner that is fair, accountable, and respectful of individual autonomy. The Executive Order sets the stage for a comprehensive approach to AI governance that not only encourages innovation but also provides robust protections for individuals and clear pathways for accountability and recourse when things go wrong, thereby fostering a trustworthy AI ecosystem that aligns with American values of justice and fairness.

Relationship to Other Instruments

The Executive Order on Ensuring a National AI Policy Framework is designed to complement and build upon existing legal and policy instruments within the United States, rather than supersede them entirely. It explicitly directs federal agencies to review and, where necessary, update their internal policies, regulations, and guidance to align with the principles and directives set forth in this order, while ensuring consistency with existing statutory authorities. For instance, the order acknowledges the foundational role of privacy laws such as the Privacy Act of 1974 and the Health Insurance Portability and Accountability Act (HIPAA), and directs agencies to ensure that AI development and deployment fully comply with these and other relevant data protection statutes. It also builds upon previous executive actions related to AI, such as earlier executive orders on maintaining American leadership in AI, by providing a more detailed and expansive framework for governance, risk management, and ethical considerations. The order serves as an overarching directive that provides a unified vision for AI policy across the federal government, aiming to harmonize disparate agency efforts and ensure a cohesive national strategy. It is intended to provide a high-level framework that will inform the development of more specific regulations and guidelines by individual agencies, ensuring a consistent approach while allowing for sector-specific nuances. This approach ensures continuity with established legal principles while adapting them to the unique challenges and opportunities presented by AI.

Furthermore, the Executive Order explicitly directs agencies to consider the implications of AI for existing civil rights laws, such as the Civil Rights Act of 1964 and the Americans with Disabilities Act (ADA), ensuring that AI systems do not perpetuate or exacerbate discrimination. It mandates that agencies assess how AI technologies interact with these laws and propose necessary adjustments to ensure equitable outcomes and protect vulnerable populations. The order also references the importance of intellectual property laws in fostering innovation and directs agencies to consider how AI impacts copyright, patent, and trade secret protections, potentially informing future legislative proposals in this area. In the realm of national security, the Executive Order builds upon existing defense and intelligence policies, directing agencies to integrate AI considerations into their strategic planning and operational frameworks, consistent with established legal authorities and oversight mechanisms. The framework is designed to be a living document, anticipating that its principles will be further elaborated and codified through subsequent legislation, agency rule-making, and international agreements. By clearly articulating its relationship to existing legal instruments, the Executive Order seeks to create a coherent and integrated regulatory landscape for AI, avoiding conflicts and leveraging established legal foundations to address emerging challenges effectively. This ensures that the national AI policy is not developed in a vacuum but is firmly rooted in the broader legal and policy traditions of the United States.

International Alignment

The Executive Order places significant emphasis on the importance of international alignment and cooperation in shaping global AI norms and standards. Recognizing that AI is a global technology with cross-border implications, the order directs the Department of State, in coordination with other relevant agencies, to actively engage with international partners, multilateral organizations, and standard-setting bodies to promote a shared vision for responsible AI development and deployment. This includes advocating for democratic values, human rights, and the rule of law in international AI discussions, ensuring that global AI governance frameworks reflect American principles. The order specifically calls for collaboration on developing common technical standards for AI safety, security, and interoperability, which can facilitate cross-border data flows and reduce regulatory fragmentation. It recognizes that a fragmented international regulatory landscape could hinder innovation and create competitive disadvantages, making international cooperation essential for fostering a robust and ethical global AI ecosystem. The Executive Order aims to position the United States as a leader in shaping international AI policy, working with allies to counter authoritarian approaches to AI governance and promote an open, secure, and reliable AI future.

Furthermore, the Executive Order directs federal agencies to participate in international forums and initiatives focused on AI, such as the Global Partnership on Artificial Intelligence (GPAI) and the OECD AI Principles, to share best practices, coordinate research efforts, and address global challenges like AI-driven disinformation and cyber threats. It emphasizes the importance of developing mutual recognition agreements for AI testing and certification, which can streamline international trade and facilitate the adoption of trustworthy AI systems across borders. The order also addresses the need for international cooperation on AI ethics, particularly concerning the responsible use of AI in military applications and the prevention of an AI arms race. It calls for diplomatic efforts to establish international norms and confidence-building measures to reduce risks associated with autonomous weapons systems and other sensitive AI technologies. By prioritizing international alignment, the Executive Order seeks to ensure that the United States' national AI policy is not developed in isolation but is part of a broader global effort to harness the benefits of AI while mitigating its risks, fostering a cooperative environment for addressing shared challenges and promoting a future where AI serves humanity's best interests worldwide.

Implementation Timeline

Sources and References