Recently, the Cyberspace Administration of China, together with multiple government departments, officially released the Interim Measures for the Management of Generative Artificial Intelligence Services (hereinafter referred to as the "Measures"), which set forth clear requirements for the labeling management of generative AI services. As a company deeply engaged in AI technology applications, we promptly organized our compliance and technical teams to conduct a thorough review of the relevant provisions, ensuring that our business practices are fully aligned with the latest regulatory requirements.
Article 12 of the Measures explicitly stipulates that organizations and individuals providing generative AI services shall, in accordance with relevant national regulations, apply explicit labels to generated images, videos, audio, and other content, and clearly inform users through the interactive interface or service agreement that such content is generated by AI. This requirement is designed to safeguard the public's right to know and prevent deep synthetic content from causing misinformation or confusion.
Explicit Labels: Applicable to all generated content provided to the public, including but not limited to text, images, audio, and video. Labels that are difficult to tamper with should be placed in prominent positions on the content.
Implicit Labels: For content that may be edited or redistributed, it is recommended to embed implicit markers such as metadata to facilitate source tracing.
Interactive Notices: During user interactions with AI services, clear notifications such as "This content is generated by AI" should be provided through interface text, pop-ups, or voice prompts.
In light of the transitional arrangements outlined in the Measures, we have initiated a comprehensive upgrade of our internal labeling system. Currently, we are advancing the following three key initiatives:
Technical Integration: Integrating labeling modules into our proprietary content generation pipeline to automatically attach explicit watermarks and implicit metadata to output content.
Process Review: Revising content publishing guidelines to require that all externally released AI-generated materials undergo a secondary compliance review to verify label completeness.
User Notification: Updating service agreements and product interfaces to add generated-content explanations at key touchpoints such as chat windows and result pages.
We recommend that clients using generative AI services pay close attention to their labeling obligations in their specific business scenarios. If your company uses AI-generated content for public dissemination or commercial purposes, please ensure compliance self-inspection of labeling is completed before publication. We also offer labeling solution consulting to help ecosystem partners quickly meet regulatory requirements.
The release of the Measures marks a new stage of standardized development for the generative AI industry. We will continue to monitor subsequent implementation rules and standards to strike a balance between product innovation and compliant operations, delivering safe and trustworthy intelligent services to users. Should you wish to learn more about specific labeling technologies or compliance details, please feel free to contact our legal and product teams.
Previous：The Party Branch Carries out Thematic Learning Activities
Next：Summer Travel Boom: Top Trending Destinations and Their Must-See Events
Scan WeChat QR code
