China Issues Interim Measures to Regulate Generative Artificial Intelligence Services


Most read

Loading Most Ready posts..


  • China introduces regulations for generative AI services, balancing innovation and compliance.
  • Obligations for service providers include data compliance, personal info protection, and content moderation.
  • Violations of the new measures may result in penalties under existing cybersecurity laws.

On July 13, 2023, the Cyberspace Administration of China (CAC) and six other Chinese government agencies announced the release of the Interim Administrative Measures for Generative Artificial Intelligence Services. These measures will take effect on August 15, 2023, marking China’s first step in regulating generative AI services. The measures aim to balance promoting industry innovation and development while ensuring the responsible and compliant use of generative AI technologies. Here are the key takeaways from the newly released measures:

Scope of the measures

The measures apply to the provision of generative AI services, encompassing the generation of texts, pictures, audio, videos, or other content for the public in China. While the term “to the public” remains undefined, it is presumed to include individual and enterprise users unless the CAC provides further guidance. Notably, internal research and generative AI technologies are exempt from these regulations if they do not involve providing services to the public.

Chinese users are not explicitly prohibited from using generative AI services overseas providers offer. However, the CAC can take necessary measures on such services if they violate Chinese laws and regulations.

Key obligations of service providers

The measures establish a comprehensive regulatory framework, imposing several compliance obligations on providers of generative AI services. Some of the key requirements include:

1. Training data compliance

Service providers must lawfully source their data and foundational models for training activities. They must ensure that their training processes do not infringe on the intellectual property rights of others. If personal information is processed during training, providers must obtain explicit consent from individuals or rely on another legal basis for processing. Additionally, steps must be taken to enhance the quality and authenticity of training data while adhering to relevant cybersecurity, data security, and personal information protection laws.

2. Data labeling and quality assurance

Service providers conducting data labeling during technology research and development must create clear, specific, and feasible labeling rules in compliance with the measures. They are also required to assess the quality of labeling and carry out spot-checks to verify labeling accuracy. Proper training must be provided to personnel responsible for labeling work.

3. Personal information protection

Apart from obtaining consent, service providers must comply with personal information protection obligations. This includes refraining from collecting unnecessary personal information and not retaining or sharing information and records that could illegally identify users. Providers must also respond promptly to user requests related to their personal information.

4. Content moderation and illegal activities

Service providers are responsible for promptly addressing illegal content on their platforms by suspending generation and transmission and removing such content from their services. If users are found using generative AI services for illegal activities, providers must issue warnings, limit their access to relevant features, and may suspend or terminate their services. Relevant records must be retained and reported to competent regulators.

5. Security assessment and algorithm filing:

Generative AI services with “public opinion attributes or the capacity for social mobilization” require security assessment conducted by the CAC and other relevant regulators. Additionally, the service providers must file their algorithms following China’s Internet Information Service Algorithmic Recommendation Management Provisions within 10 business days after the services’ launch.


Violations of the measures will be subject to penalties under various existing laws, including the Cybersecurity Law, Data Security Law, Personal Information Protection Law, and the Law on Scientific and Technological Progress. Competent regulators have the authority to request cooperation from service providers for supervision and inspection purposes, seeking explanations on training data sources, labeling rules, algorithmic mechanisms, and necessary technical and data support.

With the release of the Interim Administrative Measures for Generative Artificial Intelligence Services, China has taken a significant step in regulating generative AI technologies and services. The measures aim to balance fostering innovation in the industry while ensuring compliance with existing laws and safeguarding users’ rights and data. As the implementation date approaches, service providers operating in China must diligently adhere to these new obligations to avoid potential penalties and regulatory actions.

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Share link:

John Palmer

John Palmer is an enthusiastic crypto writer with an interest in Bitcoin, Blockchain, and technical analysis. With a focus on daily market analysis, his research helps traders and investors alike. His particular interest in digital wallets and blockchain aids his audience.

Stay on top of crypto news, get daily updates in your inbox

Related News

Subscribe to CryptoPolitan