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

The U.S. government's framework for evaluating and approving the release of frontier AI models, such as OpenAI’s "Sol," lacks transparency and clear criteria. Experts and industry insiders note that the current regulatory environment relies heavily on ad hoc discussions between tech companies and officials rather than a formalized, standardized process. Concerns persist regarding the influence of political donations and personal connections on regulatory outcomes, as well as the absence of independent, rigorous expert oversight. Critics argue that this secretive, opaque system creates uncertainty, misaligned incentives, and a lack of public trust, potentially favoring industry interests over comprehensive safety assessments.

(Source：TechCrunch)