The Trump administration is considering creating an independent watchdog to review the safety of the most advanced artificial intelligence models, following complaints from Silicon Valley that recent U.S. efforts to slow model releases have been ad-hoc and unpredictable.
Frontier AI Oversight · Under Review
A Wall-Street-Style Watchdog for the Most Powerful AI
The Trump administration is weighing an independent body — modeled on FINRA and sitting under the SEC — to vet advanced AI models before wide release, after Silicon Valley called recent case-by-case interventions "ad-hoc and unpredictable."
FINRA
The self-regulatory model for the proposed body — industry input, government supervision
30 days
Voluntary government access window for a "covered frontier model" under the June 2 order
0
Mandatory licensing or preclearance requirements — the order stopped short of hard mandates
The Proposal, Shrinking
From 90 days to 30 — access window scaled back over China competitiveness
90 days
earlier proposal (dropped)
→
30 days
June 2 executive order
Why It Shifted — The Trigger Case
Anthropic held back general availability of its Mythos Preview after finding it could discover and exploit cybersecurity vulnerabilities.
Limited access via Project Glasswing
Microsoft
Amazon
Apple
Supporters
Industry participation in setting standards could reduce friction and give companies clearer, more predictable rules of the road — if it stays light-touch.
Skeptics
A self-regulatory body may struggle to solve the underlying safety problems it is meant to address — and caution could cede competitive ground.
Still undefined: review criteria · capability thresholds that trigger scrutiny · timelines · penalties. The plan sits with the White House Chief of Staff with no formal announcement — oversight remains in flux.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…