Trump Revokes Biden's AI Executive Order

In his first week in office, Donald Trump signed an executive order dismantling the artificial intelligence policy inherited from the previous administration. The document, titled Removing Barriers to American Leadership in Artificial Intelligence and dated January 23, 2025, revokes Executive Order 14110 and replaces its approach with one explicitly focused on competitiveness and technological dominance.

The change is not cosmetic. The revoked order, signed on October 30, 2023 under the title Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, had been the cornerstone of federal AI policy. Its disappearance marks a fundamental shift in how the United States understands its regulatory role toward the technology of the moment.

What the New Order Says

The text is short and direct. Its purpose section states that the United States "has long been at the forefront of artificial intelligence (AI) innovation," driven by "the strength of our free markets, world-class research institutions, and entrepreneurial spirit." It adds a line that sets the tone for the entire document: to maintain that leadership, the country "must develop AI systems that are free from ideological bias or engineered social agendas."

The official policy the order establishes is to "sustain and enhance America's global AI dominance" in order to "promote human flourishing, economic competitiveness, and national security." The language of dominance replaces the vocabulary of guardrails and safeguards that defined the previous framework.

To turn that intent into concrete action, the document sets two deadlines:

• 180 days for the Assistant to the President for Science and Technology, the Special Advisor for AI and Crypto, and the Assistant to the President for National Security Affairs, in coordination with other senior officials, to develop and submit an AI action plan that puts the new policy into practice.
• 60 days for the Director of the Office of Management and Budget (OMB) to revise memoranda M-24-10 and M-24-18 to bring them in line with that same policy.

The order also directs the heads of federal agencies to immediately review all policies, directives, regulations, and orders adopted under Executive Order 14110, identify those that conflict with the new approach, and suspend, revise, or rescind them. Where that rollback cannot be completed right away, the text calls for granting "all available exemptions" in the meantime.

What's Being Dismantled

Executive Order 14110 was the most ambitious instrument any US administration had deployed on AI. It laid out a framework for "safe, secure, and trustworthy" development, with requirements and obligations spread across multiple federal agencies. The new order doesn't reform that framework — it revokes it outright and orders a review of everything done under its authority.

The mechanism chosen here matters. Rather than swapping the old framework for a new set of rules, the administration orders the dismantling first and leaves the definition of what replaces it for six months down the line. During that window, the federal regulatory reference point on AI is essentially emptied out, pending a document that doesn't exist yet.

The two OMB memoranda under review are the other front. These documents govern how federal agencies themselves acquire and use AI systems. Rewriting them to align with the "dominance" policy means changing the internal rules the government uses to buy and deploy this technology.

From Caution to Acceleration

The contrast between the two approaches is the key to understanding what just happened. The revoked framework started from a precautionary premise: set limits, requirements, and safeguards before accelerating. The new order flips that order of priorities, putting competitiveness and leadership ahead of safeguards, which the text barely bothers to mention.

The reference to "ideological bias" and "engineered social agendas" points to different territory altogether. It doesn't define what counts as bias or how it will be measured, but it introduces ideological neutrality of AI systems as a policy objective — an angle absent from the previous approach. How that translates into concrete criteria will depend on the action plan and the rewritten memoranda.

It's worth measuring the real scope of an executive order. It's a tool of the executive branch, not a law: it directs federal agencies and can be reversed by a later signature, as this very document just demonstrated with respect to its predecessor. Its own closing clause acknowledges as much, clarifying that the order "is not intended to, and does not, create any right or benefit" enforceable in court, and that it will be implemented "subject to the availability of appropriations." It's a statement of policy direction and an internal work schedule, not a set of obligations for companies.

What to Watch in the Coming Months

The substance of this shift hasn't been written yet. The order sets the framework and the tone, but leaves what will actually change the rules of the game to two future deliverables: the 180-day action plan and the 60-day revised OMB memoranda.

Until those texts exist, what's tangible is the withdrawal of the previous framework and the mandate for agencies to review — and, where applicable, unwind — what was done under Executive Order 14110. Everything else — how the goal of "dominance" gets defined, what development "free from ideological bias" means in practice, which obligations disappear and which, if any, replace them — remains open to whatever those deadlines produce.

The political message, though, is already on the table: the United States is formalizing its shift from AI governance oriented around safeguards to one oriented around winning the race.