The White House would not comment on whether it intends to publicly release a framework for voluntary reviews of advanced artificial intelligence models that it plans to brief companies on later today. According to reporting by Semafor, major AI firms including Anthropic, OpenAI, Google and Meta had not been shown the framework’s details ahead of the meeting.
Deadline and completion
A White House official said that former President Donald Trump signed an order in June requiring the framework to be completed by Aug. 1. The official said the framework was finished on schedule.
Content expectations and the stakes of non-disclosure
The framework is expected to reflect evaluations of advanced models that are already underway. If the document is not publicly released, companies that were not briefed could face a difficult decision about whether to opt into an ostensibly voluntary review process without having seen its full contents.
Unspecified related event mentioned in invitations
People familiar with the plans told Semafor that invitations to today’s event also referenced a related, but unspecified, event. That reference adds uncertainty about who will have access to the framework and any associated processes.
What remains unclear
Key outstanding questions include whether the White House will publish the framework and, if so, in what form; which firms or stakeholders will be able to participate in the voluntary reviews; and what implications those decisions will have for industry cooperation and regulatory practices around advanced AI.
Source: Semafor (Eleanor Mueller)
