Trump's AI Testing Plan: Limited and Vague Framework Excludes Open Models

According to Axios, the Trump administration's framework for assessing cybersecurity risks from advanced AI excludes open models outright, meaning anyone can download and inspect them. The framework explicitly states it cannot be used to restrict open models after release. It was created after President Trump signed an executive order in June requesting AI companies share frontier models with the federal government before release to address cybersecurity concerns. Companies like Anthropic, OpenAI, and Google attended a briefing on the finalized framework at the White House, but the administration won't release framework details publicly. The framework sets a 30-day grace period for government review, but it only applies to closed-source models with state-of-the-art capabilities and national security risks, without defining those terms. While voluntary, frontier labs are seeking guidance, and the ambiguity could be problematic for smaller AI providers.