Reports say White House AI review framework may stay private

In short: Multiple reports say the White House is working on a voluntary AI model review framework, and it is not planning to release it publicly.

What happened

Several news reports, citing people familiar with the discussions, say the White House has drafted a framework for evaluating advanced AI models, but does not plan to publish it for the public to read.

The process is described as voluntary. Instead of a public rollout, AI companies would review the draft in private meetings with the administration. Reports say these meetings were expected to involve the Office of the National Cyber Director.

The framework is linked to a June executive order. It is meant to help AI developers and the government decide whether a new, very capable model falls under the government’s review system. Think of it like a set of guidelines for when you need an inspection before opening a new factory, except the “factory” is an AI system.

Details about how the review would work are still unclear. Reports suggest it could focus on “frontier” models, meaning the most advanced systems available. Some coverage also suggests the government could evaluate certain models before they are released to the public, potentially up to 90 days in advance, but the exact rules and thresholds have not been shared.

The White House has not publicly committed to releasing the framework. Reuters has previously reported that officials declined to confirm details, beyond saying any policy announcement would come directly from the president.

Why it matters

If the framework stays private, the public, researchers, and smaller companies may not know what standards the government is asking major AI labs to meet. That matters because advanced AI can affect everyday life, from job tools to security risks, and people often want to understand what safety checks happen before new systems are widely used.

Source: NYTimes