On August 7, 2026, White House officials met with representatives from top AI companies such as Anthropic, OpenAI, Microsoft, Meta, Google, and Nvidia to discuss a new voluntary pre-release review process. This framework aims to establish a security risk assessment mechanism for leading “state-of-the-art” AI models in the United States under the Trump administration, marking a shift in policy focus from support to stronger regulation in response to Chinese competition. However, the specific details of this framework were only made public to the participating companies, many core concerns regarding AI security were not addressed, and neither the government’s evaluation process nor information about the participants was disclosed.
White House officials welcomed representatives of leading AI companies on Tuesday to discuss a new voluntary pre-review process. This framework will determine how the Trump administration reviews the security risks of America’s leading “leading-edge” AI models, marking a shift from supporting AI to strengthening regulation in response to Chinese competition. However, details are only available to the companies involved (including Anthropic, OpenAI, Microsoft, Meta, Google, and Nvidia). According to leaked information, this framework does not address many core concerns regarding AI security, and neither the government’s evaluations nor the participating parties are made public.