The White House is in advanced talks with major AI companies to set voluntary standards for releasing new models, with an announcement possible as soon as next week, the Financial Times reported on July 2, 2026.
Frontier AI Policy · The White House
White House Nears Voluntary Standards for Releasing New AI Models
Advanced talks with major AI companies could yield an announcement "as soon as next week." As of now there is no signed agreement — only ongoing discussions, building on the June 2, 2026 executive order on AI innovation and security.
30days
Advance access review period offered to government (voluntary)
0
Mandatory licensing or preclearance — explicitly rejected
NSA
Classified benchmarking led by, focused on cyber capabilities
Review Window: Proposed vs. Framework
The advance-access period was trimmed from an initial proposal of 90 days.
90 days
initially proposed
30 days
current framework
Framework at a Glance
Scope
Covered frontier models designated by the NSA
Assessment focus
Advanced cyber capabilities (classified benchmarks)
Provided access
Under confidentiality; early sharing with trusted partners
Enforceability
Mandatory licensing and preclearance rejected
Supportive View
OpenAI's Sam Altman was reported to say the approach strikes a balance between safety and innovation.
Critical View
Some experts warn a nominally voluntary scheme could become a de facto licensing regime, and that a 30-day review may delay releases — with questions over how trusted partners are chosen.
Still to come. The specific terms, participating companies and assessment methods have not been disclosed and will depend on forthcoming official statements.
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…