Artificial intelligence-generated synthetic content identifiers with clear visible and hidden markings

The Central Network Office developed information on relevant methods to clarify the responsibilities of those who produce synthetic content identification and dissemination services.

Open information released by the Central Network Office in March 2025 introduced the Artificial Intelligence Generation Synthetic Content Marking Scheme.The method is to divide the synthetic content identifiers into visible and hidden identifiers: the former give the user a direct sense of the text, sound or graphics, and the latter, through technical measures, write into the resulting synthetic content document data.

For publishing institutions and content platforms, the marking mechanism relates not only to multimedia materials such as pictures, audio and video, but also to source management and responsibility boundaries in content production processes.When artificial intelligence-assisted creation is used, the source of the material, the use of the tool and the record of manual editing should be kept and the necessary markings and clearances should be completed in accordance with applicable rules.

At the same time, it is emphasized that the relevant markings should not be removed, altered, forged or concealed in bad faith.This paper provides only a summary of authoritative and publicly available information, and does not judge agency-specific compliance.In the case of specific operations, the laws and regulations in force, departmental regulations, national standards and professional opinions shall prevail.

Source：Public source　View details