meet with the Trump administration

The White House is set to gather artificial intelligence firms on Tuesday to review a framework the administration recently finalized for evaluating the cybersecurity strength of their most sophisticated models, a White House source told CNBC. The meeting, which has not yet been announced publicly, will center on the voluntary program that President Donald Trump initiated in June. Representatives from Anthropic, OpenAI, and Google are anticipated to take part, according to sources familiar with the plans.

The framework was developed in response to Trump's June 2 executive order, which tasked federal officials with establishing a process for AI developers to assess whether their models qualify as "covered frontier models." Through the voluntary program, these developers can grant the government access to their models for up to 30 days prior to making them accessible to other trusted partners.

This early access could enable both the government and the tech companies to evaluate the potential of powerful models to uncover software vulnerabilities or execute complex cyberattacks.

The administration has emphasized that the early access would be used solely for evaluation purposes and not to establish mandatory federal licensing, permitting, or preclearance requirements for the development or release of new AI models. The benchmark and the threshold for determining which models should undergo review remain classified. The White House has not revealed the completed framework or the specific metrics it will employ to test participating models.