In Washington, the Trump administration has completed its voluntary framework for assessing new artificial intelligence models, as confirmed by a White House official. A meeting is scheduled for Tuesday at the White House with industry partners to discuss the framework, although specific details about its contents have not been disclosed.
Axios initially reported that the framework's finalization stems from an executive order signed by President Trump in June. This directive seeks to bolster both the security and innovative potential of AI technologies by establishing a program that encourages AI companies to voluntarily share their advanced models with the government prior to public release.
The executive order highlights the intention of the federal government to promote innovation without imposing excessively restrictive regulations. It also emphasizes the necessity of reinforcing the country's federal cybersecurity frameworks to incorporate AI technologies effectively. Additionally, it outlines a process to identify cutting-edge AI models—referred to as "frontier" models—and invites companies to allow the government access to these advanced systems for a period of up to 30 days before their release.
Concerns about the potential misuse of frontier models, particularly regarding their ability to uncover previously unnoticed software vulnerabilities in critical systems, have emerged. In response, Anthropic, a prominent AI research lab, announced in April that it would collaborate with select partners to provide its new model, Mythos, enabling them to enhance their defenses against cyber threats ahead of a wider release.
The executive order underlined the voluntary aspect of collaboration between AI companies and the federal government, assuring that it would not restrict innovators in their pursuit of technological advancement.
