Overview of Executive Order 14110
Executive Order 14110, signed on October 30, 2023, by President Joe Biden, established the U.S. government’s most comprehensive framework for regulating artificial intelligence (AI) at the time. Titled Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, the order aimed to address rapid advancements in AI—particularly generative models like ChatGPT—while mitigating risks to national security, civil liberties, and economic competitiveness. It was the 126th executive order of Biden’s presidency and the third AI-focused executive order issued by the U.S. government, following two prior orders under President Donald Trump.
The order was revoked by President Donald Trump shortly after his inauguration on January 20, 2025, as part of broader rescissions of Biden-era policies. Despite its repeal, Executive Order 14110 remains significant because its provisions influenced subsequent agency programs, standards, and regulatory approaches in the AI sector.
Key Policy Objectives and Provisions
The executive order outlined five primary policy goals to guide AI governance: promoting competition in the AI industry, protecting civil and labor rights, safeguarding consumer privacy, developing technical standards for AI-generated content, and maintaining U.S. leadership in global AI innovation. These objectives were framed in response to concerns about AI-enabled threats, including misinformation, discrimination, and national security vulnerabilities.
To achieve these goals, the order mandated several concrete actions across federal agencies. One of the most notable provisions was the requirement for major agencies—such as the Department of Homeland Security (DHS), the Department of Veterans Affairs (VA), and the National Science Foundation (NSF)—to appoint chief artificial intelligence officers. These roles were intended to centralize AI oversight within government operations, ensuring coordinated policy implementation and risk management.
The order also directed the Department of Homeland Security to develop AI-related security guidelines, including cybersecurity protocols and coordination with private-sector firms in critical infrastructure sectors like energy. Additionally, the VA was tasked with launching an AI technology competition to reduce occupational burnout among healthcare workers by deploying AI-assisted tools for routine administrative tasks.
Agency-Specific Responsibilities and Technical Standards
The National Institute of Standards and Technology (NIST), under the Department of Commerce, was assigned a critical role in expanding its AI Risk Management Framework to include specific guidance for generative AI systems. This framework was designed to help organizations assess and mitigate risks associated with AI deployment, aligning with broader efforts to standardize AI safety and reliability.
This provision aimed to address intellectual property concerns and the spread of deepfakes or synthetic media, which could undermine trust in digital information. The order also emphasized protecting consumers from AI-enabled harms, such as algorithmic discrimination or privacy violations, though it did not impose mandatory disclosures on AI training data or models.
Notably, the executive order did not include proposals for licensing advanced AI models or prohibiting high-risk applications—a gap that had been advocated by some industry leaders, including figures like Sam Altman of OpenAI. Instead, the order relied on voluntary compliance and agency-led initiatives to foster responsible AI development.
Reception and Legislative Context
The executive order received mixed reactions from political, industry, and civil society stakeholders. Democratic lawmakers, including Senator Richard Blumenthal and Representative Ted Lieu, praised the order as a step toward responsible AI innovation, though they emphasized the need for congressional legislation to strengthen regulations. Republican critics, such as Senator Ted Cruz, argued that the order created unnecessary barriers to innovation under the guise of safety.
Public opinion polls indicated broad support for the order, with 69% of voters approving its provisions, including majorities across Democratic, Republican, and independent voters. However, industry groups like the Chamber of Commerce, NetChoice, and the Software and Information Industry Association criticized the order for potentially stifling private-sector innovation, warning that regulatory overreach could hinder technological progress.
Civil rights organizations, including the American Civil Liberties Union (ACLU) and the Leadership Conference on Civil and Human Rights, welcomed provisions addressing AI fairness and discrimination but urged Congress to take further action. The order was issued amid growing legislative efforts in the 118th U.S. Congress, where lawmakers like Senate Majority Leader Chuck Schumer had pushed for comprehensive AI regulation. Despite its repeal, Executive Order 14110 set a precedent for federal engagement in AI governance, influencing later discussions on risk management and ethical deployment.
Legacy and Influence on Subsequent AI Governance
Its requirements—such as the establishment of chief AI officers and the expansion of NIST’s risk framework—were incorporated into ongoing agency initiatives. The order also reinforced the Biden administration’s AI Action Plan, which included voluntary safety commitments from major tech companies and the AI Bill of Rights, a blueprint for protecting individuals from AI-related harms.
The Defense Production Act (DPA) provisions in the order, which required reports on frontier AI model training, reflected an early attempt to monitor high-risk AI development. The order’s emphasis on collaboration between public and private sectors also became a recurring theme in subsequent AI policy discussions, particularly in addressing cybersecurity threats and critical infrastructure vulnerabilities.
Required actions
- Was: report dual-use foundation model training runs and red-team results to Commerce under the Defense Production Act - Was: agencies to adopt AI governance, NIST to develop evaluation guidance, and cloud providers to report foreign training customers