Analysis of China’s Draft Administrative Measures for Digital Virtual Human Information Services

In this analysis for Anjie Broad, author Xiangxiang Ma dives into China’s proposed regulatory framework aimed at addressing the misuse of AI-generated virtual humans that infringe personality rights such as name, image, and reputation. Released by the Cyberspace Administration of China in April 2026, the Draft Measures respond to growing concerns that AI-enabled tools—such as face-swapping and voice cloning—are being used at scale to evade traditional enforcement mechanisms. The article positions the draft as part of China’s broader effort to close legal gaps arising from rapidly advancing generative AI technologies.

The Draft Measures establish a detailed definitional framework that categorizes key actors in the digital virtual human ecosystem, including service providers, technical supporters, and end users. This tripartite structure is designed to allocate responsibilities based on each actor’s role in the value chain, while maintaining a technology-neutral approach that can accommodate future developments. The broad definition of “digital virtual human” ensures that both real-time, human-driven avatars and fully AI-generated entities fall within the scope of regulation.

Substantively, the framework focuses on rights protection and operational standards. It introduces safeguards for personality rights, personal data, and intellectual property, alongside specific protections for minors. At the same time, it imposes detailed compliance obligations on service providers, including risk monitoring systems, human and technical oversight, content labeling requirements, and mechanisms for handling user complaints. Entities with broader societal impact are also subject to algorithmic accountability rules, while enforcement provisions establish a tiered system of penalties depending on the severity of violations.

The article also situates the Draft Measures within China’s expanding AI regulatory architecture and highlights areas for refinement. These include the need for more proportionate, risk-based obligations—particularly for smaller firms—clearer alignment with existing laws such as the Civil Code, and more granular differentiation of responsibilities across actors. It further suggests that a phased implementation approach and additional regulatory guidance would improve compliance and adaptability as the technology and market continue to evolve.