Trump Administration's AI Policy Faces Internal Conflict and Shifts
Internal Trump administration disagreement and delays over AI policy, amid a push for federal vetting of advanced models sparked by Anthropic's Mythos model, are stalling action on AI safety.
Objective Facts
The Trump administration, which positioned itself as opposite to Biden by criticizing AI safety efforts and embracing anti-regulation under AI czar David Sacks, is now reportedly considering oversight for advanced AI models driven by cybersecurity concerns about Anthropic's Mythos model's ability to identify and exploit vulnerabilities. David Sacks departed his role as AI czar in March, with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent since taking more active roles in shaping AI policy. Administration disagreement and Trump's China summit are stalling federal response efforts, with early talk of safety reviews slowing as officials show they are not on the same page. National security officials want more sway in AI regulation as the White House grapples with cybersecurity threats from advanced AI models.
Left-Leaning Perspective
Groups arguing for more AI oversight have cheered the administration's recent comments, with Brad Carson, president of Americans for Responsible AI, calling the potential shift "a promising step in the right direction." Brookings Institution scholars argue the Trump administration's AI Action Plan signals a troubling shift away from safe, accountable AI toward rapid private-sector deployment, while simultaneously gutting the National Science Foundation—the agency expected to carry out the plan's mandates—with over 1,600 active grants supporting STEM and foundational AI abruptly canceled using political criteria. Commentators note that algorithmic bias and discrimination, data privacy beyond children, transparency, and environmental impacts are entirely absent from the framework despite being among the most well-documented AI risks, making the document read primarily as an industry growth strategy with limited safeguards rather than genuine oversight. The Brennan Center for Justice warns that unregulated AI in national security poses risks to privacy and civil rights, and that the administration's enforcement of "truth-seeking" requirements embeds the assumption that the administration's version of truth—excluding recognition of climate change and transgender people—produces neutral outcomes, leaving Americans with faulty technology that divides and misinforms. The Center for American Progress argues the executive order directing federal challenges to state AI laws represents an unprecedented, unconstitutional, and dangerous assertion of executive power, utilizing an "AI Litigation Task Force" and funding restrictions primarily to serve as threat and intimidation rather than legitimate governance. Democratic representatives including Doris Matsui and Ted Lieu have introduced the GUARDRAILS Act to repeal Trump's AI framework and block federal preemption of state regulation, with substantive Democratic proposals coalescing around limiting federal preemption, strengthening oversight mechanisms, and establishing safeguards against harmful AI deployment.
Right-Leaning Perspective
David Sacks, who left his formal leadership role on AI but whose worldview is "largely written into the framework," stated recently on the "All-In" podcast that he doesn't believe AI poses an existential threat and that "people are treating this like some existential threat... I don't think it is, as long as everyone does what they're supposed to do" by using AI tools to bolster security. Conservative policy analysts Neil Chilson and Adam Thierer argue that "adopting an FDA-style regulatory regime for AI would represent a shocking policy reversal by the Trump administration, and a major about-face on how America has approached software, online speech, and digital commerce," warning that Hassett's recent comments about vetting have caused significant concern. PJ Media contributor argues that dominant platforms with progressive predispositions have used their position to entrench themselves, and risks AI following the same trajectory; Anthropic CEO Dario Amodei is accused of being a vocal critic opposing the administration's decision to allow AI chip exports and comparing approved sales to "selling nuclear weapons to North Korea," which the outlet characterizes as attempting to "frame the president's approach as reckless and dangerous" and box policymakers into restrictions that benefit his company while discouraging competitors. Academic analysis in Science magazine argues the Trump administration is not removing government from AI regulation but rather concentrating governmental power at the federal level through investment, ownership, research funding, and preemption—advancing not the absence of regulation but its rearrangement through industrial policy and executive discretion—and attempting to set the terms on which AI governance will be understood globally.
Deep Dive
The Trump administration entered its second term explicitly positioned as the opposite of Biden on AI, with AI czar David Sacks embodying a strong anti-regulation ethos that rejected what Trump advisors saw as overly burdensome AI safety efforts and licensing regimes. The catalyst for internal conflict was Anthropic's announcement of its Mythos model—so powerful at identifying cybersecurity vulnerabilities that the company restricted its release, prompting administration disagreement and delays in response as officials grapple with balancing innovation against newly urgent national security concerns. The leadership vacuum created when David Sacks departed in March has given White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent more influence over AI policy, shifting the center of gravity from pure deregulation toward security-focused oversight, creating the observed internal tension. Helen Toner from Georgetown's Center for Security and Emerging Technology offered a fair assessment that Sacks' hands-off approach "was not a well-considered whole of government, really thoroughly endorsed position" but rather reflected one person's strong views dominating the room. A former Trump White House official acknowledged to The Hill that "the flip-flopping nature of the administration's tech response signal that there is no clear direction or leader driving the agenda" and that internal whiplash on AI undermines action on actual risks. What each side gets right: Progressive critics correctly identify the administration's contradictions—it claims to oppose "coercion" of tech companies on content while simultaneously demanding removal of "woke" elements and threatening federal litigation against states; conservatives correctly note that the Mythos response does mark a dramatic tactical reversal from the administration's deregulatory platform. What each side misses: Progressives understate the genuine national security dimension of cybersecurity vulnerability automation; conservatives understate how the administration's approach remains fundamentally hostile to state-level consumer protections and transparency requirements even amid tactical shifts on federal vetting. The unresolved question is who ultimately controls AI policy direction—security agencies seeking more regulatory power, or Trump himself, as the White House has stated any formal policy announcement will come directly from the president. The administration may wait on the outcome of Trump's China summit before making final decisions, with industry anxiously awaiting new guidance on frontier AI model releases. The outcome will determine whether recent safety rhetoric signals genuine policy pivot or tactical positioning ahead of midterm elections.