Trump administration close to finalizing voluntary AI review framework

According to a report by The Information, three people familiar with the matter disclosed that the Trump administration is on the verge of finalizing a voluntary framework mandating AI companies to submit their most advanced models to the government prior to public release. The sources noted that the White House Office of the National Cyber Director circulated a draft framework to OpenAI, Anthropic, and Google (GOOG.O) roughly two weeks ago, with all three firms submitting their respective revisions. An AI-related executive order signed by Trump in early June set August 1 as the deadline for developing this framework, which is designed to establish a more standardized review process for powerful new AI models capable of rapidly detecting cybersecurity vulnerabilities. As the deadline draws near, some AI companies have raised concerns in meetings with White House agencies over how the framework will define frontier models and whether it will differentiate between closed-source and open-source models, two sources said. Another outstanding question is the actual level of voluntariness of the framework in practice, particularly after the White House imposed export controls on Anthropic. The Trump administration maintains that compliance with the framework is not mandatory. (Jinshi)