Original title: 16 global technology companies and institutions sign the commitment of cutting-edge AI security

China News Online, May 23 (Song Yusheng, a reporter from China News Finance and Economics) learned from Zhipu AI that 16 technology companies and institutions have recently signed a security commitment for cutting-edge AI, and will strive to ensure the security of their most advanced AI models, and implement responsible governance to achieve openness and transparency.

The 16 companies involved in the commitment include enterprises in the United States, France, China, South Korea and the United Arab Emirates, including Amazon, Microsoft, Meta, Google, OpenAI, xAI in the United States, Mistral AI in France, Samsung Electronics and Naver in South Korea, G42 and Technology Innovation Institute (TII) in the United Arab Emirates, and Zhipu AI in China.

The key points of cutting-edge AI security commitment include: ensuring a responsible governance structure and transparency of cutting-edge AI security; Based on the AI security framework, responsibly explain how to measure the risks of cutting-edge AI models; Establish a clear process for risk mitigation mechanism of cutting-edge AI security model.

This security commitment is based on the Bletchley Declaration issued by the first AI Security Summit last November, and aims to promote global cooperation in AI security.