Measures for the Administration of Digital Virtual Human Information Services (Draft for Comment)
📑 Legal hierarchy: Level 3 · Departmental rule (draft-for-comment stage, not yet effective) | Issuance: Cyberspace Administration of China (single-ministry draft for comment) | Released: 2026-04-03 | Character: hard law (once effective)
⚠️ Hierarchy note: This instrument is currently a draft for public comment, not yet in its final form. When finalized, it will be a departmental rule. The final version may be jointly issued by multiple ministries (see the 2026-04-10 five-ministry joint Interim Measures for the Administration of AI Anthropomorphic Interaction Services). See Index of Chinese Rules.
🆕 2026-04-03 Draft for comment. Drafted by CAC pursuant to CSL / DSL / PIPL / the Measures for the Administration of Internet Information Services. Together with the 2026-04-10 Interim Measures for the Administration of AI Anthropomorphic Interaction Services, it forms a sister-rule pair: the Anthropomorphic Measures focus on the “interaction-behavior layer,” while this instrument focuses on the “digital-image layer.” Archived copy at
public/archives/china/digital-virtual-human-services-draft-2026-04-03.html.
Chinese Summary
Section titled “Chinese Summary”The Cyberspace Administration of China drafted the Measures for the Administration of Digital Virtual Human Information Services (Draft for Comment) pursuant to the Cybersecurity Law, Data Security Law, Personal Information Protection Law, and Measures for the Administration of Internet Information Services, and opened it for public comment on 2026-04-03.
Core Definition (Draft for Comment)
Section titled “Core Definition (Draft for Comment)”数字虚拟人：存在于非物理世界，利用图形学、数字图像处理或者人工智能等技术，借助真人驱动或者计算驱动，模拟人类外貌，具备声音、行为、交互能力或者性格等特征的虚拟数字形象。
— English gloss — “Digital virtual human” means a virtual digital image existing in a non-physical world, produced with technologies such as computer graphics, digital-image processing, or artificial intelligence; driven by a real person or by computation; simulating human appearance; and possessing features such as voice, behavior, interactive capability, or personality.
Scope includes:
- Virtual anchors, AI news presenters;
- Virtual idols, virtual spokespersons;
- AI-driven anthropomorphic customer-service / companions;
- Digital doubles driven by real-person motion capture.
Known Key Provisions (from 2026-04-03 commentary)
Section titled “Known Key Provisions (from 2026-04-03 commentary)”Identity notice and conspicuous labeling
Section titled “Identity notice and conspicuous labeling”自数字虚拟人服务开始，需在其展示区域全程进行显著提示标识。
— English gloss — From the start of a digital virtual human service, a conspicuous prompt/label shall be maintained throughout the display area.
The use of likeness or voice highly similar to a specific natural person in a way that misleads the audience is prohibited.
Article 7 · Likeness / Voice Protection
Section titled “Article 7 · Likeness / Voice Protection”Without consent, services that can identify a specific natural person via a digital virtual human shall not be provided.
Sensitive personal information requires consent; upon withdrawal, relevant information shall be deleted.
Article 10 · Protection of Minors (mirroring Article 14 of the Anthropomorphic Measures)
Section titled “Article 10 · Protection of Minors (mirroring Article 14 of the Anthropomorphic Measures)”- Inducing addiction of minors is prohibited;
- No “virtual relatives / virtual lovers” services shall be provided.
Note: nearly identical to Article 14 of the Anthropomorphic Interaction Services Measures, reflecting consistency between the two rules on minors’ protection.
Article 11 · Protection of Heroes’ and Martyrs’ Images
Section titled “Article 11 · Protection of Heroes’ and Martyrs’ Images”Shall not be used for commercial purposes; interfaces with the Law on the Protection of Heroes and Martyrs.
Rights of Drivers (the real person behind the virtual human)
Section titled “Rights of Drivers (the real person behind the virtual human)”Protects the “drivers” (actors, anchors, voice-over artists) behind real-person-driven digital virtual humans:
- Personal-information protection;
- Freedom to choose employment (preventing long-term lock-in or sale of a person’s digital double by an employer).
Article 20 · Complaints and Reports Mechanism
Section titled “Article 20 · Complaints and Reports Mechanism”Platforms shall establish convenient user complaint and reporting mechanisms.
Relationship with the Anthropomorphic Interaction Services Measures
Section titled “Relationship with the Anthropomorphic Interaction Services Measures”A dual-rule system (CAC’s two-step move in April 2026):
| Dimension | Anthropomorphic Interaction Services Measures (issued 2026-04-10) | Digital Virtual Human Measures (2026-04-03 draft for comment) |
|---|---|---|
| Issuer | Five-ministry joint | CAC alone (final text may be joint) |
| Focus | Interaction-behavior layer (AI companionship, language/emotion interaction) | Digital-image layer (visual presentation, likeness and voice) |
| Overlap in scope | Virtual companionship, AI partners | Virtual anchors, digital doubles, virtual idols |
| Minors clause | Article 14 — bans virtual companions | Article 10 — same text |
| Dedicated clause for older persons | Article 15 present | To be confirmed against final text |
| Dedicated clause on likeness / voice | No strong dedicated clause | Articles 7, 11 — strong dedicated clause |
| Driver protection | Not present | Unique to this rule — protects real-person anchors |
Structurally this is a complementary coverage of “interaction side (anthropomorphism) + image side (digital virtual human).”
Relationship with Upstream / Adjacent Rules
Section titled “Relationship with Upstream / Adjacent Rules”- Deep Synthesis Provisions (2023): the “image generation” part of digital virtual humans is already covered; these Measures deepen the rule at the service layer.
- Generative AI Interim Measures (2023): the underlying AI used in the interaction part remains subject to them.
- Labeling Measures (2025): synthesized content produced by digital virtual humans is still subject to labeling.
- PIPL: treats likeness / voice as sensitive information.
- Law on the Protection of Heroes and Martyrs: the upstream authority for Article 11.
Source Text and Archival Copies
Section titled “Source Text and Archival Copies”| Source | Link |
|---|---|
| Chinese draft for comment (CAC) | cac.gov.cn/2026-04/03/c_1776952992709096.htm |
| Chinese (archived copy) | archives/china/digital-virtual-human-services-draft-2026-04-03.html |
| English | China Law Translate |
Version History
Section titled “Version History”| Date | Event |
|---|---|
| 2026-04-03 | Draft for public comment released |
| ~2026-05 (estimated) | End of comment period |
| Second half of 2026 (estimated) | Official release |