China promotes human-centered AI governance

China’s National Technical Committee 260 on Cybersecurity published the Ethics-Safety Guidelines for Artificial Intelligence Applications 1.0 on Tuesday, outlining a framework for ethical and safe AI development. The document, released at the 2026 China Internet Civilization Conference in Nanning, Guangxi, emphasizes human-centered AI governance, prioritizing human control over AI systems and ensuring AI serves 'the common well-being of humanity.' The guideline establishes key principles, including risk management, openness, privacy protection, and trustworthy AI systems. It mandates human judgment, supervision, and intervention at critical stages to prevent AI from replacing human decision-making entirely. For users, it advises moderation in AI use, warning against over-reliance on AI tools for emotional support or interpersonal interactions. The document also addresses AI-generated content, urging stronger risk governance to combat 'information cocoons, cognitive misleading, and cognitive degradation.' Additionally, it encourages an open-source innovation ecosystem, promoting transparency in AI models, tools, and evaluation benchmarks while enhancing security in open-source platforms. The guideline applies to AI development, service provision, and usage, serving as a reference for organizations and individuals engaged in AI-related activities. It positions China’s approach as principle-based, aiming to balance innovation with ethical and safety standards in AI applications.