Huawei: Chinese AI too weak to grasp rogue-AI risks

Huawei's rotating chairman Eric Xu said Chinese AI developers may not be advanced enough to perceive the safety risks reported by leading U.S. firms, urging them to accelerate development while balancing innovation against risk. Xu made the remarks at Huawei's annual Connect conference in Shanghai, noting that U.S. providers with massive computing power may uniquely understand their models' capabilities and associated dangers. His comments follow intensified U.S. alarm over AI safety, with OpenAI pledging to disclose unexpected model behavior and Anthropic's CEO calling for slower development. Chinese researchers and state media have criticized such calls as attempts to preserve U.S. technological leadership. Beijing treats AI as manageable and is developing mandatory standards, while Huawei forecasts autonomous agents will account for over 90% of global AI processing traffic by 2035.