This is part of a series in the course of Covington blogs on policies, executive decrees and other actions of the Trump administration. The first blog summed up the key measures taken during the first weeks of the Trump administration, including the revocation of President Biden in 2023 Executive command 14110 on “safe, secure and trustworthy development and use of AI” and the release of President Trump Executive command 14179 On “the abolition of obstacles to American leadership in artificial intelligence” (“AI EO”). This blog describes actions on the AI taken by the Trump administration in February 2025.
People from the White House Request for information plan for AI Action Plan
On February 6, the Office of the Science and Technological Policy of the White House (“OSTP”) published a Request information (“RFI”) requesting the public’s contribution to the content which should be in the Action Plan of the White House. The RFI marks the first important step of the Trump administration in the implementation of the very broad objectives of AI EO of January 2025, which requires that the assistant of the president for science and technology Michael Kratsios, the White House Ai and the Tsar Crypto David Sacks, and the national security adviser Michael Waltz to develop an “action plan” for the world policy of promoting improvement in America) competitiveness and national security. “RFI stipulates that the AI action plan” will define the priority political actions necessary to maintain and improve the domination of the American AI, and to ensure that unnecessarily heavy requirements do not hinder the innovation of the private sector “.
More specifically, the RFI requests public comments on the “highest political actions” which should be included in the AI action plan and encourages respondents to recommend “concrete” actions necessary to solve AI policy problems. While noting that the answers can “resolve any subject of relevant AI policy”, the RFI provides 20 subjects for potential entry. These subjects are general and do not include specific questions or areas where a particular contribution is necessary. Subjects include: equipment and chips, data centers, energy consumption and efficiency, model development and open source, confidentiality and data security, technical and security standards, national security and defense, intellectual property, supply and export controls. As of March 13, more than 325 comments on the AI action plan were submitted. The public commentary period ends on March 15, 2025. Under the OE, the Action Plan finalized AI must be submitted to the President in mid-October 2025.
Vice-president JD Vance describes the AI political priorities at the top of the AI action in Paris at the top of the action
On February 11, vice-president JD Vance gave a scanning political discourse to government, industry and civil society leaders at the IA 2025 action summit in Paris. In his remarks, the vice-president underlined the “countless revolutionary applications” of the AI to “make us more productive, more prosperous and more free” and described four key political priorities of the AI for the Trump administration: (1) Ensure that “American technology of the AI continues to be the gold stallion”; (2) Encourage “Pro-Corporate Policy of AI” instead of “excessive regulation of the AI sector”; (3) Ensure that AI “remains free from ideological biases” and is not “co -opted in an authoritarian censorship tool”; and (4) the maintenance of a “pro-work-and-worker growth plan for AI so that it can be a powerful tool for the creation of jobs in the United States”. Vice-president Vance also said that the next Action Plan of the White House AI “would avoid () a regulatory diet too precautionary while ensuring that all Americans benefit from technology and its transformer potential”.
In particular, the vice-president Vance underlined the disagreement of the Trump administration with EU technological regulations, warning that “the future of AI will not be won by making hands on security” and by calling for “international regulatory regimes that promote () the creation of AI technology, rather than strangulation ()”. Following Vice-President’s speech, the United States and the United Kingdom refused to connect to the top of AI action Joint declaration On “inclusive and lasting artificial intelligence for people and the planet”.
Nist is looking for a public comment on the Cyber AI profile
On February 14, the National Institute of Standards and Technology (“NIST”) announcement The creation of a new “community profile” to provide risk management advice linked to the “cybersecurity of AI systems, AI compatible cyber attacks and AI cyberfense cyberfense” (the “Cyber IA profile”). Nist also published a concept paper On the cyber ai profile, which stats that organizations “Need Risk Management Approaches That Support the Realities of Advancements in A Ai Use to Position them for Defending Against Aitabled Cyber Offense by Adversaries and Taking Advantage of Ai-Enabled Cyber Defense Capabilities,” and notes that the cyber Nist cybersecurity framework 2.0published in February 2024, or the NIST AI Risk Management FrameworkPublished in January 2023. The Document Concept notes that the NIST has made a certain number of observations related to cybersecurity around AI, including that there is no framework on how companies should use AI to protect systems and that AI does not fundamentally change the way organizations must approach cybersecurity. The Document Concept also lists twelve questions for public comments, in particular if “the design and failures of the implementation of the AI” should be included in the Cyber IA profile and if the Cyber IA profile should tackle “the relationship between cybersecurity and confidentiality of the AI”. Public comments on the Cyber AI profile is due before March 14, 2025. NIST plans to hold a Cyber and AI workshop To discuss the document concept on April 3, 2025.
President Trump issues a memorandum on the first investment policy in America
On February 21, President Trump published a Memorandum Presidential National Security (“NSPM”) on the America First Investment Policy. The NSPM establishes the American policy of “preserving (ING) an open investment environment to help guarantee that artificial intelligence and other emerging technologies of the future are built, created and cultivated here in the United States”, while also announcing foreign investment restrictions in response to “predative investment actors and technology-acquisition of RPC and other foreign adversaries or threat”. Among other things, the NSPM provides for an “accelerated” process for American investments of American allies and partners in advanced technologies, accelerated environmental examinations for investments of more than $ 1 billion, new rules for stopping American investments in Chinese AI and other technologies and Chinese investments in “restrict access to the United States”. The NSPM orders the Treasury Department and the CFIUS to take measures under the International Law on Economic Powers, the Defense Production Act and other statutes to implement the NSPM.
Congress and States react to the Deepseek and US-China AI breed
State legislators and American federals have continued to answer In the US increased competition in the AI, after the release of the R1 AI model in Deepseek in January. In the Congress, members of the two parties called for the prohibition of using it from Deepseek AI models on American government devices. On February 7, representatives Josh Gottheimer (D-NJ-5) and Darin Lahood (R-I-16) presented the No Deepseek law on the government (HR 1121), which would order the White House management and budget office (“omb”) require that federal agencies prohibit the use of the Deepseek AI application, or any successor application or service, on agency information technologies. Similarly, on February 20, Senator Tom Cotton (R-AR) sent a letter To Camb director Russell Vought, calling on the OMB to “immediately prohibit the American government’s services and agencies from using and accessing Deepseek” or other AI tools developed in China, citing national security problems linked to Deepsek data security practices. At the state level, officials from New York, Virginia, Iowa and Pennsylvania issuing new prohibitions on The installation or use of Deepseek AI models on state government devices and networks or by government entrepreneurs.
Likewise, American legislators have also published bipartite calls for new sales checks in order to limit the development of Chinese AI for national security reasons. On February 3, the senses. Josh Hawley (R-Mi) and Elizabeth Warren (D-MA) sent a letter To the Secretary of Commerce, Howard Lunick urging the Commerce Department to “strengthen our export controls on the PRC” in response to the R1 and V3 models of Deepseek, including by strengthening the AI diffusion rule of the Biden administration and by restoring the export of H20 or the equivalent semi-clonding chips and other essential components. On February 7, representative Mark Green (R-TN-7) presented China Technology Transfer Control Act (HR 1122), which would order the president to control the export to China of any “national technology or national IP”, including AI technology, and to prohibit foreign transactions to sell or buy AI technology to or from China.