The administration will meet with leading AI companies to review a voluntary framework for assessing advanced models.
The administration will host a meeting on Tuesday with major artificial intelligence companies, including Google, OpenAI, and Anthropic, to discuss a newly completed framework for reviewing the cybersecurity capabilities of advanced models. This voluntary framework was established by an executive order from President Trump in June, which directs federal officials to create a process for developers to determine if models qualify as 'covered frontier models.' Under this program, participating developers can provide the government with early access to their models for up to 30 days before public release. This allows the government and technology companies to evaluate how powerful models might be used to identify software vulnerabilities or conduct cyberattacks. The administration seeks a light-touch regulatory approach, ensuring the program does not establish mandatory federal licensing or preclearance requirements. While the framework and specific benchmarks remain classified for now, the administration announced that the process is complete and discussions with industry partners are underway to address safety and national security risks.