OpenAI, Anthropic, Google, and Meta will meet with White House officials on Tuesday to review a completed voluntary framework that grants the government 30 days of access to frontier AI models before public release, according to CNBC and CNN.
The framework allows the four companies to submit models for government testing without legal mandate, making participation voluntary rather than regulatory. The meeting comes as the EU's AI Act enforcement took effect and follows a week in which more than 1,200 senior staffers across major AI labs signed an open letter calling on Washington to help pace frontier AI development.
The voluntary nature of the framework distinguishes it from the regulatory posture the EU has taken. Under the EU's AI Act, which entered enforcement on August 3, companies face mandatory compliance with tiered rules based on model risk classification. The U.S. framework instead relies on industry cooperation, with no statutory force behind the 30-day access period.
The timing of the White House meeting comes after President Biden's 2023 executive order on AI safety set broad goals around model testing and safety. This framework provides mechanics through the voluntary agreement with the four largest frontier model developers in the market.
The 30-day pre-release window gives government agencies time to test models for potential harms before they reach the public. OpenAI, Anthropic, Google, and Meta collectively account for the majority of frontier model releases in the market. No other AI companies are confirmed to have signed the framework or been invited to the Tuesday meeting.
The White House has not disclosed whether it plans to expand the framework to other AI developers or make participation a condition of federal contracting. The framework's success depends on whether the four signatory companies treat the 30-day window as binding and whether the government builds sufficient testing capacity to review models in that timeframe.
