How did the government decide OpenAI’s frontier model was safe to release?

The United States government has not disclosed the specific criteria or internal discussions used to determine that OpenAI and Anthropic’s frontier models were safe for public release. While these companies have voluntarily agreed to safety evaluations, the process remains opaque to the public and independent researchers. This lack of transparency raises questions about how federal agencies oversee the deployment of powerful AI systems and what standards define acceptable risk before such technology reaches the market.