OpenAI founder agrees with Anthropic CEO on frontier model control

OpenAI founder Sam Altman (奥尔特曼) stated on September 13 that he fully agrees with Anthropic (Anthropic) CEO Dario Amodei (达里奥) regarding the necessity of controlling the progress of frontier AI models. Altman confirmed that OpenAI plans to grant independent third-party evaluators system access similar to that of internal employees to enhance AI safety. Detailed control protocols and evaluation frameworks are currently under development, with further updates expected soon.