What to Know About the New Trump Administration Executive Order on Artificial Intelligence

Yesterday, the Trump administration debuted its broad approach to artificial intelligence (AI) with a short executive order (EO) titled “Removing Barriers to American Leadership in Artificial Intelligence.” The policy goal, as described in Section 2 of the document is clear: “[T]o sustain and enhance America’s global AI dominance in order to promote human flourishing, economic competitiveness, and national security.” It directs a series of actions that build on the day-one repeal of Biden Administration Executive Order 14110, titled “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence,” which the Trump administration argues excessively burdened American AI companies in ways that would undermine U.S. AI technology leadership, or reflected ideological bias or social agendas. The multiple AI actions by the White House during the first week of the administration reinforces the centrality that AI, as a general-purpose technology rapidly advancing and impacting the lives of every American, will play across the Trump administration. Here are key takeaways. 

First, the Trump EO directs a 180-day review of all implementing actions that occurred due to the Biden administration EO. The tasks outlined in the Biden administration EO were essentially completed prior to the end of the administration, so the repeal of the EO itself had little substantive impact. Now, the Trump administration will review the implementing actions that occurred due to the EO. 

Second, one of the most important actions the Biden administration AI EO triggered was Office of Management and Budget (OMB) Memoranda 24-10, Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence. OMB M-24-10 laid out extensive guidance surrounding how federal departments and agencies should approach AI adoption and implementation, especially concerning reducing risks from AI. It was followed by AI acquisition guidance in October 2024. Departments and agencies then released implementation plans showing how they planned to comply. Section 5 of the Trump AI EO directs an updating on those memos within sixty days. It is in the changes to those memos, and those implementation plans, that will lead to either small or large shifts in how the executive branch practically approaches the adoption of AI. 

Third, the Biden AI EO and OMB M-24-10 only partially regulated national security agencies like the Defense Department and the Central Intelligence Agency. In October 2024, the Biden administration released a subsequent memorandum titled “Memorandum on Advancing the United States’ Leadership in Artificial Intelligence: Harnessing Artificial Intelligence to Fulfill National Security Objectives; and Fostering the Safety, Security, and Trustworthiness of Artificial Intelligence.” The AI NSM reflected the Biden administration’s commitment to accelerating the adoption of trustworthy AI by national security agencies, including provisions like mandating pilot projects with frontier AI by national security agencies. It was mandated by the Biden AI EO to govern how national security agencies approach AI, so it will therefore also require re-evaluation by the Trump administration under the terms of the new EO. 

The practical impact of the Trump administration AI EO is not yet clear and may not be clear until the end of the 180-day review period and announcements of further AI policy changes. One possible pathway is that the Trump administration may decide to keep Biden-era AI policy actions that were designed to support U.S. leadership in AI (the policy goal in Section 2 of the Trump AI EO), while stripping out parts of them that they view as burdening either private sector AI innovation or federal government adoption of AI. For example, the Trump Administration could decide to keep the AI NSM in large part, given that that document was itself designed to promote U.S. national security leadership in AI, but cut regulatory or reporting requirements. A second pathway is that the Trump administration might repeal all of the key Biden AI actions and start afresh, though this might lead to regulatory uncertainty within the executive branch for longer. Finally, the Trump administration might repeal those Biden AI policies and not replace them, which would lead to more department and agency-level variation in how they approach AI adoption, as opposed to an approach broadly applicable to all departments and agencies, like unified AI procurement guidelines. 

Finally, the EO leaves unaddressed the future of the AI Safety Institute (AISI). The AISI was a key institution, based in the Department of Commerce, created during the Biden Administration to work with the private sector to accelerate trustworthy AI innovation, track and plan for advances at the AI frontier, and ensure that releases of commercial AI models do not create opportunities for misuse, such as making it easier for potential adversaries to design biological weapons. Recent research suggests the AISI needs further strengthening to enable American AI leadership and should be the focal point for public-private cooperation on frontier AI and cooperation with U.S. allies and partners on frontier AI.