The White House does not plan to make public its new framework for evaluating advanced artificial intelligence (AI) models, according to a report published Tuesday by Axios.

This voluntary framework carries global implications for AI safety, yet its specific details will only be provided to companies actively involved in the process.

On Tuesday, the White House hosted a series of meetings with AI sector representatives to review the newly finalized framework.

While the details of these discussions remain confidential, sources indicate the talks addressed open-source models and included staff from tech giant Nvidia.

Reasons Behind the Decision

According to Axios, keeping the framework private means that external companies, policymakers, researchers, and foreign U.S. allies will not know how the Administration plans to implement one of its core AI policies.

This gathering occurred shortly after Nvidia Chief Executive Officer and open-source advocate Jensen Huang visited Washington last week to meet with Commerce Secretary Howard Lutnick and President Donald Trump.

The secrecy surrounding the AI framework coincides with a period in which the industry faces significant cyberattacks alongside rapid advancements by Chinese AI developers.

The framework is expected to set conditions allowing companies to grant the government access to certain AI models up to 30 days prior to their public release.

However, the Administration will keep the evaluation process for advanced model cybersecurity capabilities classified, in accordance with the executive order signed by President Trump in June—explaining in part why it does not intend to publish the full framework.

These rules will encompass critical areas including confidentiality, cybersecurity, insider threats, intellectual property, and non-disclosure requirements.