A letter signed by Steve Bannon and about 60 allies, including Humans First leaders and church figures, urges President Trump to require government testing and approval for frontier AI systems before public release.

The letter also criticizes major AI companies and executives for insufficient accountability, arguing against allowing elites to experiment on the public without safeguards.

Humans First argues that private AI developers should not have unchecked power to release unvetted systems, calling for an executive order mandating mandatory testing, evaluation, vetting, and government approval.

Overall, the episode highlights a growing divide over whether AI should be treated primarily as an economic/geopolitical asset or a technology requiring stricter federal controls, potentially pressuring the administration to rethink its hands-off stance.

The appeal signals a political split within Trump-aligned circles, contrasting with the White House’s generally light regulatory approach and preference for voluntary standards to maintain competitive edge.

Current U.S. AI policy context includes Trump’s AI action plan and executive orders aimed at accelerating innovation, data-center development, and export promotion, with ongoing bipartisan concerns about competition with China and the regulatory landscape.

Overall tension exists between accelerating AI innovation and implementing robust safeguards, with no clear evidence of a unified federal strategy yet.

Administration officials have discussed stronger oversight options but have not endorsed mandatory pre-release approvals for AI models, indicating potential regulatory policy debates ahead.

The White House currently favors a light regulatory touch and opposes broad government approval of AI models, reflecting a hands-off approach to the AI race and state-level regulations.

The push comes amid a relatively hands-off Trump administration stance on AI regulation, following a December executive order aimed at reducing regulatory burdens in the name of national and economic security.

Previously, the administration designated Anthropic as a supply chain risk after its refusal to limit the use of its technology for certain purposes; Anthropic later sued the DoD, and Trump has claimed there have been “very good talks” with Anthropic about potential future Pentagon work.

Policy and expert context: Daniel Schiff from Purdue notes the coalition has funding and reach to influence both state and local levels, though it remains unclear how much impact the letter will have on policy.