A conservative-led US group has proposed a far-reaching AI policy platform, including independent testing of large AI companies, public reporting of AI incidents and tighter export controls. The agenda goes well beyond the Trump administration’s current light-touch approach, Semafor reports, but these are proposals, not rules in force.
Watch Desk analysis
What happened
Humans First, a group advocating stronger AI safeguards, is due to release its “America First AI Policy” platform on Friday, according to a memo shared with Semafor. Its proposals include restricting exports to China and other adversaries, requiring independent testing of large AI companies, prohibiting advanced practices such as recursive self-improvement, and requiring public reporting of AI incidents.
The group also proposes holding frontier AI developers liable for reckless conduct, tighter rules for AI used by children, a ban on federal use of cameras such as Flock, and a pause on H-1B visas. Those last measures extend beyond AI regulation narrowly defined. Semafor says the framework goes further than proposals from some Republicans in Congress and the administration’s present approach.
Why it matters
This is a policy platform from an advocacy group, not legislation or an official government plan. Still, it puts several consequential questions on the table at once: who should test powerful AI systems, what companies should disclose when things go wrong, and where the state should draw limits on development and deployment.
The package also shows how AI policy can become a vehicle for wider arguments about immigration, surveillance and corporate power. That makes the debate harder to file under one neat label, and easier for individual proposals to get lost in the pile.
Our read
The useful move is to separate the proposals and judge each on its own merits. Independent testing and incident reporting deserve a serious debate; a broad package should not get a free pass simply because some of its safeguards sound sensible. Nor should every idea be waved away because it comes wrapped in partisan branding. Policy is not a tasting menu, but readers should still be allowed to ask what is actually being served.
What to watch
- Whether the platform leads to legislation or formal proposals from elected officials.
- How the group would define independent testing, reportable incidents and reckless conduct.
- Whether the proposals gain traction among Republicans or shape the wider debate ahead of the midterms.
Discussion spark: Should AI rules be built around specific safeguards such as independent testing and incident reporting, or is a broad political platform the better way to force action?
Sources and evidence
- Conservative AI safety group proposes export controls, visa restrictions | Semafor (16 October 2026, 08:56 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.