The White House will host a meeting with representatives of artificial intelligence companies to analyze a framework for testing cybersecurity.

The White House will hold a meeting with representatives of major companies in the field of artificial intelligence, such as Anthropic, OpenAI, and Google, to discuss a new voluntary framework for testing the cybersecurity capabilities of advanced AI models.

This initiative was generated by an executive order signed by President Donald Trump in June, which allows companies to provide the government access to AI models for up to 30 days before they are publicly released. The goal is to assess potential risks, including identifying vulnerabilities and preventing sophisticated cyberattacks.

Testing will be coordinated by the Department of the Treasury, NSA, and CISA, and the technical criteria will remain classified. The administration emphasizes that this mechanism is voluntary, without the intention of imposing a mandatory licensing system. The initiative comes in the context of concerns regarding the security of AI systems, following recent incidents of platform compromises.