The White House briefed leading technology firms about a new framework on Tuesday. This framework aims to assess potential cybersecurity risks in cutting-edge AI systems. However, the administration does not plan to make the entire framework public.
According to officials, certain AI systems will be exempt from government review. Specifically, free tools referred to as “open weight” models will not undergo the same scrutiny. The focus will primarily be on assessing the latest technology from major U.S. AI developers, including companies such as OpenAI and Anthropic.
This approach allows the government to prioritize evaluating more advanced models that might pose greater risks. The move indicates a targeted strategy to ensure cybersecurity during the proliferation of AI technologies.