Trump Administration to Expand AI Model Oversight, Open-Source Models to Face Federal Reviews

ChainCatcher reported that the Trump administration plans to revise its artificial intelligence guidelines and widen government supervision of AI models. The White House has prepared an AI framework requiring U.S. laboratories to receive federal safety testing before public release of their models. According to the report, the framework is expected to include open-source models in the coming months. Open-source models will be subject to testing only after their capabilities reach a level comparable to Anthropic's Mythos-class models or OpenAI's GPT-5.6. The report did not disclose additional details, including how capability levels will be assessed or when the revised guidelines will be formally released. The change means that as open-source AI advances, developers behind the most capable open models would need to clear the same federal safety gate that the framework sets for commercial labs. The framework, as described, gives the federal government a pre-release review role for advanced AI models.