White House Mulls AI Vetting Regime
Steve Bannon and more than 60 other Trump allies sent a letter urging Trump to require government vetting of "potentially dangerous" frontier AI models before release.
Objective Facts
The Trump administration is considering an executive order to create a working group on artificial intelligence, with proposals including a government review process for new models, according to reporting by the New York Times citing US officials. The proposal has been prompted by Anthropic's Mythos model, which demonstrated capabilities for finding network vulnerabilities and posed cybersecurity risks. National Economic Council Director Kevin Hassett said the process would be similar to FDA drug approval, stating models should be "released in the wild after they've been proven safe, just like an FDA drug". The White House has worked on several draft executive orders, but which proposals reach the president's desk remains debated. Most recently, Steve Bannon and over 60 Trump allies signed a letter urging Trump to implement mandatory testing and government approval of frontier AI systems, with the letter organized by the conservative group Humans First.
Left-Leaning Perspective
TechPolicy.Press reported the White House is weighing an executive order to create a working group with government access to new AI models for vetting, potentially involving NSA, the National Cyber Director, and the Director of National Intelligence. The outlet noted this represented a departure from the administration's initial deregulatory stance. TechPolicy.Press characterized the vetting proposal as "a meaningful shift from an administration that spent its first year dismantling Biden-era AI safety frameworks," suggesting public concern may have finally registered with officials. However, the publication argued that "merely reviewing a model without consequence does not equate with meaningful oversight," and warned that "a working group co-designed with the companies being reviewed would not be any better" than relying on company discretion. The Daily Signal reported that Democrats have shown increased interest in negotiating on the Trump administration's National Framework on AI since news of possible vetting orders broke, with an AI industry source saying the potential executive orders could help Trump secure Congressional votes. This suggests some progressive lawmakers view limited vetting as a compromise position worth engaging on, though formal Democratic statements on the specific vetting regime remain limited in available reporting.
Right-Leaning Perspective
Kevin Hassett rejected the characterization of the vetting order as creating an "FDA for algorithms," arguing instead that "rather than stand up a brand-new federal body, Hassett has argued for leveraging existing government frameworks to conduct these assessments". Hassett made a geopolitical argument that "excessive regulation could drive AI innovation overseas, specifically to China," concerns echoed in broader policy discussions. Jennifer Huddleston and Juan Londoño with the libertarian Cato Institute stated that "requiring pre-launch approval was criticized as heavy-handed and anticompetitive when included in the Biden administration's executive order on AI," implying the same criticism should apply to Trump's proposed regime. The American Enterprise Institute called the proposal "a stunning reversal by the White House, which has championed AI development and criticized the sometimes heavy-handed regulations enacted by the European Union and some states. It would also be a mistake." Axios reported that inside the White House, "the prevailing view is the opposite: that America will win the AI race by keeping regulation light and knocking down most state-level AI laws," with even "administration officials who support testing and evaluating models" having "backed away from the idea that the government should approve them".
Deep Dive
The White House vetting regime proposal emerged in early May 2026 when the New York Times reported the Trump administration was considering an executive order creating a working group on artificial intelligence with proposals for government review of new models before release. The catalyst was Anthropic's Mythos model, which the company disclosed could find network vulnerabilities and posed cybersecurity risks, prompting the administration to reconsider its hands-off AI approach. The substantive disagreement centers on whether government pre-release vetting represents genuine safeguards or counterproductive regulation. Progressive critics correctly note that AI companies face incentives against rigorous safety testing and that independent verification matters in high-stakes industries—they point out pharmaceutical companies conduct clinical trials under FDA frameworks, yet equivalent safeguards do not exist in AI. Conservative economists counter that licensing requirements impose high compliance costs favoring established incumbents, reduce the number of market players, and reduce innovation by requiring standardization to meet regulatory requirements. Some analysts note that Mythos's signature capability—finding vulnerabilities at scale—has been achievable using cheaper models, questioning whether Anthropic's dramatic framing justifies extraordinary oversight measures. The administration itself appears divided. There are internal disagreements about vetting strength, with some officials preferring a light touch while others want aggressive model vetting. Hassett rejected creating an "FDA for algorithms," arguing for leveraging existing frameworks rather than new bureaucracy. Yet Axios reported the White House prevailing view remains deregulatory, with even vetting supporters backing away from government approval of models. Most recently, Steve Bannon and 60 Trump allies pushed back against this stance, with Bannon telling Axios that "mandatory testing and government approval" is a "must" for AI.