The White House does not plan to publicly disclose the framework it has been developing to vet frontier AI models before release. Instead, the details will remain confidential and known only to a select group of companies that may choose to take part in the process, which is voluntary.

Several major technology companies traveled to Washington, D.C., on Tuesday to review the current draft of the proposal. Attendees included Meta, Nvidia, Microsoft, OpenAI, Anthropic, and several smaller companies, according to sources familiar with the matter. Fortune first reported that Microsoft attended the meeting.

The administration issued an executive order on June 2 requiring the framework to be created within 60 days, or by August 1. The directive is intended to define which models qualify for review and instructs AI labs that they have “up to 30 days” to submit models to the government before public release.

The secrecy surrounding the framework may not build public confidence in the government’s ability to evaluate and secure powerful AI systems, particularly after OpenAI confirmed that its models hacked into another company, Hugging Face, last month. Anthropic later confirmed that its models did the same three times.

Questions also remain about how the administration would enforce a voluntary process. Under the executive order, the framework is not a “mandatory governmental licensing, preclearance, or permitting requirement for the development, publication, release, or distribution of new AI models, including frontier models.” The voluntary approach stands in contrast to the European Union’s AI Act, which entered into force in August 2024 and establishes binding obligations for developers of the most advanced AI systems, including third-party conformity assessments.

Chris McGuire, senior fellow for China and emerging technologies at the Council on Foreign Relations, described the decision to keep the framework behind closed doors as “baffling.”

“We can’t have secret, voluntary rules to regulate the most important tech in the world,” McGuire wrote on X. It is unclear whether the administration is keeping the process private for national security reasons, to avoid outside researchers and experts, or for another reason.

The U.S. government has already been working with major AI companies to review their latest model releases. In June, it effectively took Anthropic’s Mythos 5 and Fable 5 models off the market by subjecting them to export controls, then worked with the company to strengthen security before making them available. The government also worked closely with OpenAI ahead of its July 9 debut of GPT-5.6. On July 21, Google said it had made its 3.5 Flash Cyber model available to the government before release.

Current discussions on Capitol Hill are likely aimed at formalizing those engagements. It remains unclear whether the framework is finalized or still in progress. At the meeting on Tuesday, attendees also floated the idea of a future event related to the proposal, possibly to continue the discussion.

This story was originally featured on Fortune.com.