Measures for Labeling Artificial Intelligence-Generated and Synthesized Content (Draft Translation)

 

Measures for Labeling Artificial Intelligence-Generated and Synthesized Content (Draft Translation)

Article 1: In order to promote the healthy development of artificial intelligence (AI), regulate the labeling of AI-generated and synthesized content, protect the legitimate rights and interests of citizens, legal persons, and other organizations, and safeguard the public interest, these measures are formulated in accordance with the Cybersecurity Law of the People’s Republic of China, the Regulations on the Management of Algorithm Recommendation in Internet Information Services, the Regulations on the Management of Deep Synthesis in Internet Information Services, the Interim Measures for the Management of Generative Artificial Intelligence Services, and other laws, administrative regulations, and departmental rules.

Article 2: Internet information service providers (hereinafter referred to as “service providers”) that engage in the labeling of AI-generated and synthesized content in accordance with the Regulations on the Management of Algorithm Recommendation in Internet Information Services, the Regulations on the Management of Deep Synthesis in Internet Information Services, and the Interim Measures for the Management of Generative Artificial Intelligence Services shall be subject to these measures.

Article 3: AI-generated and synthesized content refers to text, image, audio, video, virtual scene and other information generated and synthesized using artificial intelligence technology.

The labels of AI-generated and synthesized content include explicit and implicit labels.

Explicit label refers to the label added to the AI-generated and synthesized content or interactive scene interface, which is presented in the form of text, sound, graphics, etc. and can be clearly perceived by users.

Implicit label refers to the label added in the file data of the AI-generated and synthesized content by technical measures, which is not easy to be clearly perceived by users.

Article 4: If the AI-generated and synthesized services provided by service providers fall under the circumstances specified in Article 17, Paragraph 1 of the Regulations on the Management of Deep Synthesis in Internet Information Services, explicit labels must be added to the AI-generated and synthesized content in accordance with the following requirements:

1.                For text, explicit labels such as text prompts or universal symbols should be added at the beginning, end, or appropriate positions in the middle of the content, or be added around the text or within the interactive interface.

2.                 For audio, explicit labels such as voice prompts or rhythmic audio cues should be added at the beginning, end, or appropriate positions in the middle of the content, or be placed within the interactive interface.

3.                 For images, explicit labels should be added in appropriate positions.

4.                 For videos, explicit labels should be added at the initial frame and appropriate locations around the video playback interface. It may also be added at appropriate positions in the middle or at the end of the video.

5.                When presenting virtual scenes, explicit labels should be added at an appropriate position in the initial frame. They may also be added at appropriate positions during the continuous service process of the virtual scene.

6.                 Other AI-generated and synthesized service scenarios should include explicit labels based on their specific application characteristics.

When service providers offer functions such as downloading, copying, or exporting AI-generated materials, they must ensure that the files contain explicit labels that meet the required standards.

Article 5: Service providers shall, in accordance with Article 16 of the Regulations on the Management of Deep Synthesis in Internet Information Services, add implicit labels in the file metadata of AI-generated and synthesized content. These implicit labels shall include attributes of the AI-generated content, the name of the service provider, or other production-related information such as code or content identification numbers.

Service providers are encouraged to embed implicit labels in the form of digital watermarks or similar forms in the AI-generated and synthesized content.

File metadata refers to descriptive information embedded in the file header using a specific encoding format, which is used to record details such as the file’s source, attributes, and intended use.

Article 6: Providers of online information content dissemination services shall take the following measures to regulate the dissemination of AI-generated and synthesized content:

1. Verify whether the file metadata contains implicit labels. If the metadata explicitly indicates that the content is AI-generated or synthesized, appropriate measures shall be taken to add prominent labels around the published content to clearly inform the public that the content is AI-generated.

2. If implicit labels are not found in the file metadata, but the user declares the content as AI-generated, appropriate measures shall be taken to add prominent labels around the published content to remind the public that the content may be AI-generated.

3. If neither implicit label is found in the file metadata nor the user declares the content as AI-generated, but the service provider detects explicit labels or other signs of AI-generation and synthesis, the content shall be identified as suspected AI-generated and synthesized content. Appropriate measures shall be taken to add prominent labels around the published content to alert the public that the content is suspected to be AI-generated and synthesized.

4. Provide necessary labeling functions and remind users to proactively declare whether their published content contains AI-generated elements.

For situations described in Paragraphs 1 to 3 above, the service providers shall add information to the file metadata such as the attributes of AI-generated content, the name of the dissemination platform, or the dissemination factors information such as code or content identification numbers.

Article 7: When reviewing an application for listing or launch, an internet application distribution platform should require the internet application service providers to indicate whether it offers AI-generated and synthesized services. If the service provider does offer AI-generated and synthesized services, the distribution platform should verify the relevant materials related to the identification of the AI-generated and synthesized content.

Article 8: Service providers should explicitly specify the methods, formats, and other standard details of AI-generated and synthesized content labels in the user service agreement. They should also remind users to carefully read and understand the relevant labeling management requirements.

Article 9: If a user requests AI-generated and synthesized content without explicit labeling, the service provider may offer such content after clearly defining the user’s labeling obligations and usage responsibilities in the user service agreement. Additionally, the service provider must retain relevant logs, including information about the users, for no less than six months in accordance with the law.

Article 10: Users who publish AI-generated and synthesized content using internet information content dissemination services must proactively declare and use the labeling functions provided by the service providers.

No organization or individual shall maliciously delete, alter, forge, or conceal the labels of AI-generated and synthesized content as stipulated in these measures, and shall not provide tools or services to others for the purpose of carrying out the aforementioned malicious actions, nor shall they use improper labeling methods to damage the legitimate rights and interests of others.

Article 11: Service providers engaging in identification activities must also comply with the requirements of relevant laws, administrative regulations, departmental rules, and mandatory national standards.

Article 12: When service providers carry out procedures such as algorithm filing and security assessments, they must provide materials related to the labeling of AI-generated and synthesized content in accordance with these measures. They should also strengthen the sharing of labeling information to support and assist in preventing and combating related illegal activities.

Article 13: If any offense violates these measures, responsibility shall be pursued by relevant authorities of cyberspace administration, telecommunications, public security, and radio and television in accordance with relevant laws, administrative regulations, and departmental rules.

Article 14: These measures shall take effect on September 1, 2025.

 

Issue Date: March 7, 2025

Authorities:

Cyberspace Administration of China

Ministry of Industry and Information Technology of the People’s Republic of China

Ministry of Public Security of the People’s Republic of China

National Radio and Television Administration

 

 

 

 

Kommentarer

Send en kommentar

Populære opslag fra denne blog

Arrangement om AI 21/2-2025 kl. 16-18 - Skrivehandlingen, sprogene, oversættelserne og den kunstige intelligens

Billede
  Dansk PEN Sprognetværk og Dansk Oversætterforbund inviterer til arrangement om:   Billedet er genereret i Copilot med min prompt De fleste af os blev blæst omkuld i efteråret 2022, da kunstig intelligens blev folkeeje i form af ChatGPT, generativ AI og forskellige billedprogrammer. 2023 gik med at teste programmerne, 2024 med at skrive politikpapirer, holdningsdokumenter, afstandtagen eller begejstring. I dag er generativ kunstig intelligens ikke længere en ny teknologi, i dag er kunstig intelligens almindeligt i brug og umulig at undgå, hvis du tænker for din computer, din mobil eller GPS’en i din bil. Spørgsmålet vi stiller på debatarrangementet, er derfor:   Hvordan håndterer vi kunstig intelligens i vores arbejde: Skrivning, sprog, oversættelse, udgivervirksomhed?   Til at svare på spørgsmålet har vi sammensat et panel, der består af: Morten Visby, formand for Dansk Forfatterforening Juliane Wammen, formand for Dansk Oversætterforbund Nira G...
Læs mere

Censur, satire og 100 års kamp for ytringsfrihed

Billede
Dansk PEN 100 år og ICORN i København Parallelt med ICORNs årsmøde i København – ICORN er den internationale forening, der hjælper forfulgte forfattere med at finde et nyt, midlertidigt hjem – fejrede Dansk PEN 100 års jubilæum.  Dansk PEN er dén forening, der arbejder for litteratur, poesi, sproglig diversitet, ytringsfrihed, og mod hadefulde ytringer. Det skete – og sker – med en række arrangementer, der fokuserede på litteratur og kunstnerisk ytringsfrihed. Kreative ytringer Micol Hebron på Storm Museum ”Museum censurerer udstilling om censur: "Vi er ikke villige til at risikere livet for en udstilling." Museet Storm åbner dørene for offentligheden til 'Den censurerede udstilling om censur'. Over 60 satiriske værker vil blive vist, men Muhammed-tegninger og video af en trevlet koran vil ikke være iblandt: "Vi er ikke villige til at risikere livet for en udstilling," siger kurator”. Teksten er fra Altinget. Det er godt gjort, at man skal gøre sig den slag...
Læs mere

Generative artificial intelligence, writers and translators

FIT and PEN Just as the Danish Association of Certified Translators Translators' Association, DACTI,  is a member of the International Federation of Translators and Interpreters - FIT, Danish PEN is a member of PEN International. I am active in both, at national level and at international level The two international organizations have overlapping focus, still different profiles. The shared focus is language and the right to express oneself: language, authorship, translation, interpretation, indigenous languages, minoritized languages, the status of languages, freedom of expression. My ultimate goal is to pave the way for a cooperation agreement between FIT and PEN International.  I am the  Vice Chair of FIT's "External Relations Standing Committee" and a  FIT's board member. In PEN International, I am active in the "Translation and Linguistic Rights Committee" (TLRC). Generative AI  Given the shared focus, members of both organizations are naturally also...
Læs mere