White House Prepares Voluntary AI Security Review Framework for Closed Models

The White House is developing a framework to assess the security risks posed by artificial intelligence systems. The review process will be voluntary in nature, meaning AI developers will not be legally required to participate. The framework is specifically designed to cover closed-source AI models, which keep their underlying code proprietary. Open-source models that publicly share their code will be excluded from the review process. The initiative reflects growing government concern over potential security threats stemming from advanced AI technologies.