Framework

The White House is meeting with leading AI companies to discuss a new framework for assessing advanced AI model cybersecurity. Based on a June executive order, this voluntary program allows developers to submit frontier models for government evaluation up to 30 days before release. This aims to collaboratively identify potential vulnerabilities and weaponization risks, involving agencies like the Treasury and NSA in classified benchmarking. The initiative seeks to balance innovation with national security in the rapidly evolving AI landscape.