AI Regulation Tracker / Consultation
China Opens Second Comment Round on Rewritten Internet Information Services Rules With a First-Ever Smart Information Services Chapter
Regulatory summary: The Measures for the Administration of Internet Information Services (Revised Draft for Comment) 互联网信息服务管理办法(修订草案征求意见稿), reopened for a second round of public comment by the Cyberspace Administration of China (CAC) on July 3, 2026, is a proposed full rewrite that grows the current rules to six chapters and 94 articles. It.
China's cyberspace regulator has reopened public comment on a full rewrite of its Internet Information Services Management Measures, expanding the 2000-era rules to six chapters and 94 articles and adding a dedicated section that puts AI applications, algorithmic recommendation, synthetic-content labeling, and AI agents under named duties. It is a draft for comment, not law in force.
Key takeaways
- The draft replaces the current Internet Information Services rules with a far larger framework of six chapters and 94 articles and, for the first time, carves out a dedicated Smart Information Services section. That section names AI-specific duties that were previously spread across separate rules on algorithmic recommendation, deep synthesis, and generative AI, and pulls them into the core internet information service regime. It also adds a serious-breach credit blacklist and clarifies coordination among the cyberspace, telecom, and public security authorities.
- Product and compliance leads at any company offering AI-driven content generation, recommendation, or agent services to users in China; platform operators and large internet platforms; gig and platform-labor operators using algorithmic scheduling; and legal and policy teams tracking Chinese AI and content rules. Foreign providers reaching Chinese users through local entities fall in scope through those entities.
- Status: Reopened for a second round of public comment on July 3, 2026.
- Pull the Smart Information Services section, map each duty in Articles 55 to 64 against your current product, list the gaps in disclosure, labeling, opt-out, worker treatment, and agent safety, and decide whether to file a comment before August 2, 2026 while the text can still move.
| Date | Jurisdiction | Rule | Affected professionals | Status or effective date |
|---|---|---|---|---|
| 2026-07-09 | China | The draft replaces the current Internet Information Services rules with a far larger framework of six chapters and 94 articles and, for the first time, carves out a dedicated Smart Information Services section. That section names AI-specific duties that were previously spread across separate rules on algorithmic recommendation, deep synthesis, and generative AI, and pulls them into the core internet information service regime. It also adds a serious-breach credit blacklist and clarifies coordination among the cyberspace, telecom, and public security authorities. | Product and compliance leads at any company offering AI-driven content generation, recommendation, or agent services to users in China; platform operators and large internet platforms; gig and platform-labor operators using algorithmic scheduling; and legal and policy teams tracking Chinese AI and content rules. Foreign providers reaching Chinese users through local entities fall in scope through those entities. | Reopened for a second round of public comment on July 3, 2026. Draft text only, subject to change before any final issuance. |
Frequently Asked Questions
Is this measure in force now?
No. It is a revised draft in its second round of public comment as of July 3, 2026, with comments due by August 2, 2026. It would become binding only after the text is finalized and formally issued.
What is the Smart Information Services section?
A new dedicated section of the draft, roughly Articles 55 to 64, that sets duties for providers using AI to deliver internet information services, covering transparency, content labeling, recommendation, worker treatment, and AI agents.
What does Article 56 require?
It requires a smart information service provider to publicly disclose the basic principles, purpose, main operating mechanism, and training-data sources of its technology, following fairness, openness, transparency, and good-faith principles.
Does the draft require labeling of AI-generated content?
Yes. Providers must label AI-generated and synthetic content in line with national rules, and no one may maliciously delete, tamper with, forge, or hide those labels.
How can a company respond before the deadline?
The CAC notice invites comments by email to law@cac.gov.cn or by mail to its cyberspace rule-of-law bureau in Beijing, with a deadline of August 2, 2026.
Internal links
Sponsored Training
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.