Chinese AI regulation proposal.md

Translated from http://www.cac.gov.cn/2023-04/11/c_1682854275475410.htm

In order to promote the healthy development and standardized application of generative artificial intelligence (AI) technology, in accordance with the "Cybersecurity Law of the People's Republic of China" and other laws and regulations, the National Internet Information Office has drafted the "Regulations on the Management of Generative AI Services (Draft for Soliciting Opinions)" and is now soliciting feedback from the public. The public can provide feedback through the following channels and methods:

1. Log on to the China Government Legal Information Network of the Ministry of Justice of the People's Republic of China (www.moj.gov.cn, www.chinalaw.gov.cn), and enter the "Legislative Opinion Collection" section in the main menu of the homepage to provide opinions.

2. Send feedback by email to: wajscy@cac.gov.cn.

3. Send feedback by mail to: Network Security Coordination Bureau, National Internet Information Office, 15 Fucheng Road, Haidian District, Beijing 100048, and indicate "Draft for Soliciting Opinions on the Regulations on the Management of Generative AI Services" on the envelope.

The deadline for feedback is May 10, 2023.

Attachment: Regulations on the Management of Generative AI Services (Draft for Soliciting Opinions)

National Internet Information Office

---

Article 1: In order to promote the healthy development and standardized application of generative AI, these Regulations are formulated in accordance with the "Cybersecurity Law of the People's Republic of China", "Data Security Law of the People's Republic of China", "Personal Information Protection Law of the People's Republic of China", and other laws and administrative regulations.

Article 2: These Regulations apply to the development and use of generative AI products and services provided to the public within the territory of the People's Republic of China.

Generative AI, as referred to in these Regulations, refers to the technology that generates text, images, sound, video, code, and other content based on algorithms, models, and rules.

Article 3: The state supports the independent innovation, promotion, and international cooperation of AI algorithms, frameworks, and other basic technologies, and encourages the use of secure and trustworthy software, tools, computing and data resources.

Article 4: Providers of generative AI products or services shall comply with legal requirements, respect social ethics and public order, and meet the following requirements:

1. The content generated by generative AI shall reflect the core socialist values, and shall not contain content that subverts the state power, undermines the socialist system, incites separatism or undermines national unity, promotes terrorism or extremism, promotes ethnic hatred or discrimination, violence, obscene or pornographic information, false information, or content that may disrupt economic and social order.

2. Measures shall be taken in algorithm design, training data selection, model generation and optimization, and service provision to prevent discrimination based on race, ethnicity, religion, nationality, region, gender, age, profession, and other factors.

3. Respect intellectual property rights and business ethics, and shall not engage in unfair competition using algorithms, data, platforms, and other advantages.

4. The content generated by generative AI shall be true and accurate, and measures shall be taken to prevent the generation of false information.

5. Respect the legitimate interests of others, prevent harm to the physical and mental health of others, infringe upon the right of portrait, reputation, and personal privacy, and infringe upon intellectual property rights. It is prohibited to illegally obtain, disclose, or use personal information and privacy, commercial secrets.

Article 5: Organizations and individuals (hereinafter referred to as "providers") that use generative artificial intelligence products to provide chat, text, image, sound generation services, including supporting others to generate text, images, sounds, etc. through providing programmable interfaces, bear the responsibility of the content producer of the generated product; if personal information is involved, they shall assume the legal responsibility of the personal information processor and fulfill the obligation of personal information protection.

Article 6: Before providing services to the public using generative artificial intelligence products, they shall declare their security assessment to the National Internet Information Office in accordance with the "Internet Information Service Safety Assessment Provisions with Public Opinion Attributes or Social Mobilization Capabilities," and fulfill the algorithm filing and change, cancellation filing procedures in accordance with the "Regulations on the Management of Algorithm Recommendations for Internet Information Services."

Article 7: Providers shall be responsible for the legality of the pre-training data and optimized training data sources of the generative artificial intelligence products.

The pre-training and optimized training data used for generative artificial intelligence products shall meet the following requirements:

1. Comply with the requirements of the "Cybersecurity Law of the People's Republic of China" and other laws and regulations;

2. Do not contain content that infringes intellectual property rights;

3. If the data contains personal information, it shall be obtained with the consent of the personal information subject or in accordance with other circumstances stipulated by laws and administrative regulations;

4. Ensure the authenticity, accuracy, objectivity, and diversity of the data;

5. Other regulatory requirements of the National Internet Information Office for generative artificial intelligence services.

Article 8: When using manual annotation in the development of generative artificial intelligence products, providers shall formulate annotation rules that comply with the requirements of these Measures, and provide necessary training to the annotators, and sample and verify the correctness of the annotated content.

Article 9: Providers of generative artificial intelligence services shall require users to provide true identity information in accordance with the "Cybersecurity Law of the People's Republic of China."

Article 10: Providers shall clearly and publicly indicate the applicable audience, occasion, and purpose of their services, and take appropriate measures to prevent users from relying excessively or becoming addicted to the generated content.

Article 11: Providers shall be responsible for protecting users' input information and usage records during the service provision process. They shall not illegally retain input information that can infer the user's identity, create portraits based on user input information and usage, or provide user input information to others. If laws and regulations have other provisions, such provisions shall be followed.

Article 12: Providers shall not generate content with discriminatory content based on the user's race, nationality, gender, etc.

Article 13: The provider should establish a mechanism for receiving and processing user complaints, and promptly handle requests from individuals for correction, deletion, or blocking of their personal information. When it discovers or learns that generated text, images, sound, video, etc. infringe upon the portrait rights, reputation rights, personal privacy, commercial secrets of others, or do not comply with the requirements of these Measures, the provider should take measures to stop the generation and prevent further harm.

Article 14: The provider should provide secure, stable, and continuous services throughout the life cycle, ensuring that users can use them normally.

Article 15: For generated content that is found to be non-compliant with these Measures during operation or reported by users, in addition to measures such as content filtering, the provider should prevent further generation within 3 months by means such as model optimization training.

Article 16: The provider should label generated images, videos, and other content in accordance with the "Management Regulations for Deep Synthesis of Internet Information Services."

Article 17: The provider should provide necessary information that can affect user trust and choice according to the requirements of the National Internet Information Department and relevant competent authorities, including the description of the sources, scale, type, and quality of pre-training and optimization training data, manual annotation rules, scale and type of manual annotation data, basic algorithm and technical system, etc.

Article 18: The provider should guide users to scientifically recognize and rationally use the content generated by AI, not using generated content to damage the image, reputation, and other legitimate rights and interests of others, nor engage in commercial hype or improper marketing. Users have the right to report to the Internet Information Department or relevant competent authorities when they discover that the generated content does not comply with these Measures.

Article 19: If the provider discovers that users violate laws and regulations or engage in behaviors that violate commercial ethics and social morality while using AI products, including engaging in online hype, malicious posting and commenting, producing spam emails, and writing malicious software, the provider should suspend or terminate the service.

Article 20: If the provider violates the provisions of these Measures, the National Internet Information Department and relevant competent authorities shall punish them in accordance with the "Cybersecurity Law of the People's Republic of China," "Data Security Law of the People's Republic of China," "Personal Information Protection Law of the People's Republic of China," and other laws and administrative regulations. If there are no relevant laws or administrative regulations, the National Internet Information Department and relevant competent authorities shall give a warning, criticize publicly, order rectification within a time limit; if the provider refuses to rectify or the circumstances are serious, they shall order the suspension or termination of the provision of services using AI, and impose a fine of not less than 10,000 yuan but not more than 100,000 yuan. If it constitutes a violation of public security management, administrative penalties for public security management shall be given in accordance with the law; if it constitutes a crime, criminal responsibility shall be investigated in accordance with the law.

Article 21: These Measures shall be implemented as of [date] in 2023.

Context: https://news.ycombinator.com/item?id=35523011