White House Meets OpenAI, Google on AI Security Framework

Tech companies like OpenAI, Anthropic, and Google recently met in Washington to discuss the administration's new framework for reviewing advanced AI models. This framework aims to evaluate the cyber capabilities of these models. The White House indicated that the federal government plans to review only certain types of AI models from specific companies for potential security risks. So far, they do not plan to review open-source AI models, though this could change. An executive order in early June directed officials to create a benchmarking process for AI models' advanced cyber capabilities. This order requires companies to share their AI models with the government for up to 30 days before public release. While described as "voluntary," the implications for companies that don't comply remain unclear. This new oversight marks a potential shift from the administration's previous light-touch regulatory approach to AI. This development could impact how new AI models are released and adopted in the future.