White House to host AI companies Tuesday to review a new voluntary model-testing framework, CNBC reports

The Trump administration will host representatives from artificial intelligence companies at the White House on Tuesday to review a new framework for testing advanced AI models for cybersecurity capabilities, CNBC reported. The report said the government is working toward a structured process that can be used to evaluate how advanced models could affect the security of networks and systems, particularly as AI capabilities expand.

CNBC said President Donald Trump’s June executive order directed federal officials to develop a process to evaluate the cybersecurity capabilities of advanced AI models. The White House has not been cited in the provided record as publishing the framework or the meeting agenda, and an official posting confirming Tuesday’s event and its scope was not included in the supplied materials.

While the specific model-testing framework described by CNBC is not confirmed in the provided evidence through a White House document or Federal Register notice, the administration has previously issued cybersecurity-focused executive-branch materials. A White House fact sheet published June 22 said President Donald J. Trump signed an executive order related to safeguarding the nation against advanced cryptographic attacks, describing efforts to protect sensitive data and critical infrastructure. That document does not specifically describe an AI model-testing framework, but it provides context for the administration’s stated cybersecurity priorities.

The Tuesday meeting is described as a review with AI companies, suggesting a voluntary approach rather than a rulemaking process in the immediate term. Under current federal procedures, a legally binding requirement would typically require additional steps, such as a notice-and-comment rule or a formal directive to agencies, or publication of an executive order’s operative provisions. Until an official framework text is published, the extent of any obligations or enforcement tools for companies remains unclear.

The administration’s stated rationale, as characterized by CNBC, centers on evaluating cybersecurity capabilities linked to advanced AI models. That focus raises practical questions for companies about how tests are designed, which benchmarks are used, how results are documented, and whether any information is shared with the government. CNBC’s report did not provide detailed methodological requirements in the supplied packet.

If the White House framework is ultimately published or incorporated into agency guidance, companies could face additional compliance work, including internal testing pipelines and documentation processes tied to cybersecurity risk. The event timing also suggests the administration may be aiming to produce a usable framework without waiting for the slower pace of formal rulemaking.

For now, companies and affected agencies will be watching for an official confirmation of the meeting and the framework text, including any references to timelines, participating departments, and the specific cybersecurity evaluation criteria described to industry.