Guideline calls for human-centric AI

China’s National Technical Committee 260 on Cybersecurity published the 'Ethics-Safety Guidelines for Artificial Intelligence Applications 1.0' on Tuesday during the 2026 China Internet Civilization Conference in Nanning, Guangxi. The guideline, a principle-based technical document, outlines ethical and safety standards for AI development, service provision, and usage, targeting organizations and individuals engaged in AI-related activities. Drafted by leading Chinese institutions and companies—including Tsinghua University, the China Electronics Standardization Institute, Alibaba Group, Huawei Technologies, and DeepSeek—the document establishes key principles such as reasonable risk control, openness, privacy protection, controllability, and agile governance. It also addresses emerging challenges like AI-driven job displacement, decision-making responsibility, and emotional dependence, which Fan Kefeng, vice-president of the China Electronics Standardization Institute, described as growing concerns in expanding AI applications. The guideline insists AI must serve 'the common well-being of humanity' and prioritize human leadership, mandating mechanisms for human control, emergency response, and intervention at critical stages. It advises users to treat AI as a tool for assistance rather than a replacement for real-life interactions, warning against excessive reliance or addiction. For AI-generated content, the document calls for stronger risk governance to prevent 'information cocoons, cognitive misleading, and degradation.' Beyond ethical and social risks, the guideline also addresses ecological concerns, noting AI’s resource and energy consumption. Zhang Linghan, director of the Institute of AI Law and Governance at China University of Political Science and Law, emphasized that the framework provides a 'Chinese approach' to mitigating risks as AI integrates deeper into daily life and reshapes human-machine relations. The document encourages an open-source innovation ecosystem while enhancing its security capabilities.