US Pushes Voluntary AI Model Standards Ahead of Week’s End

The US government is closing in on voluntary standards for AI companies releasing new models. An announcement could come within days.

These standards aim to set clear benchmarks and timelines for advanced AI models. They also plan to define who can access these models, both inside the US and abroad. The move builds on a June executive order from the previous administration.

That executive order required AI developers to give the government early access to frontier models. But it stopped short of mandatory controls, retreating from an earlier draft that called for longer review periods. Instead, the approach relies on voluntary cooperation without licensing demands.

The White House’s real concern isn’t consumer harm. It’s the risk that advanced AI could be used by military or intelligence agencies in adversary countries like China or Russia. The new framework would let officials designate a model as a “covered frontier model” and negotiate access terms before release.

Major AI players are already in talks. Google reportedly discusses an advanced coding model under development. Microsoft, xAI, and others agreed earlier this year to let the government test their models pre-launch. Meta remains the last major holdout on pre-release security reviews.

OpenAI has faced pressure to slow down model releases while security checks catch up. CEO Sam Altman told Congress he prefers funding for testing infrastructure over a formal approval regime. This voluntary framework aligns with that preference, allowing companies to decline participation without penalty.

Several US agencies—including the Treasury, NSA, Cybersecurity and Infrastructure Security Agency, and the National Institute of Standards and Technology—have 60 days from signing to build a classified benchmarking process. Early August looks like the earliest the framework can firm up.

If finalized soon, this could mark the first consistent government standard for AI model release. It balances Washington’s strategic concerns with industry’s desire to avoid heavy-handed regulation. The next few days will show if voluntary cooperation holds or if tougher rules are imminent.