National Information Security Standardization Technical Committee, Practice Guidelines for Cybersecurity Standards—Methods for Labeling Generative Artificial Intelligence Service Content

全国信息安全标准化技术委员会秘书处网络安全标准实践指南——生成式人工智能服务内容标识方法

September 28, 2023 | BY

Susan Mok

Practice guidelines are issued specifying labeling of AI content

Issued: August 25, 2023

Main contents: In the display area of AI-generated content, prompt text shall be continuously displayed at the bottom of the display area or below the area where the user inputs information, or a prominent watermark containing prompt text shall be uniformly added to the background of the display area. The prompt text shall, at minimum, contain such information as "generated by artificial intelligence" or "generated by AI" (Article 3.1).

Where the provision of a service changes from being provided by a natural person to being provided by artificial intelligence and the same is likely to cause confusion among users, labeling thereof shall be accomplished by way of prompt text or prompting voice and such prompt text or prompting voice shall, at minimum, contain such information as "the service is provided to you by artificial intelligence" or "the service is provided to you by AI" (Article 3.5).

This premium content is reserved for
China Law & Practice Subscribers.

  • A database of over 3,000 essential documents including key PRC legislation translated into English
  • A choice of newsletters to alert you to changes affecting your business including sector specific updates
  • Premium access to the mobile optimized site for timely analysis that guides you through China's ever-changing business environment
For enterprise-wide or corporate enquiries, please contact our experienced Sales Professionals at +44 (0)203 868 7546 or [email protected]