DeepMind presents International AI Model Review Body Mask and Altmann support

On 17 July, CEO DeepMind Hathabis proposed the establishment of an international independent body to conduct a unified test and review of the front-line AI model prior to its release, and plans to go to Washington next week to discuss the relevant programme with the United States Government. According to Hasabis, the advanced cyber-attack capability of the Mythos model under Anthropic highlights the need to strengthen security controls in AI. The programme proposes that the independent expert be responsible for the review, that the head AI company provide funding, that the model be tested for up to 30 days, and that, if necessary, coordination be given to the development of cutting-edge models for the industry. Chief Executive Officers Ortmann and Mask of OpenAI have publicly expressed their support。