US Gov't Wants AI Labs to Share Models 90 Days Before Release

The U.S. government is asking advanced AI laboratories to share their models 90 days before public release. This proposal comes from a draft White House executive order. It creates a voluntary framework for reviewing advanced AI models. What's interesting is this move signals a more assertive federal approach to AI governance. The order has two main parts. One focuses on cybersecurity, aiming to protect Pentagon systems and encourage information sharing between AI companies and federal agencies. The second part deals with what are called "covered frontier models," establishing a process to define and review these powerful AI systems before they are released. The 90-day timeline is significant in the fast-paced world of AI development. While the framework is currently voluntary, history shows these frameworks can become less voluntary over time. This development shows increasing federal attention to AI oversight, balancing innovation with national security risks.