Searchlight AI
Mainland China

China has implemented several regulations to ensure the ethical and safe development of AI.

The Cybersecurity Standards Practice Guide - Guidelines for Preventing Ethical and Safety Risks in Artificial Intelligence《网络安全标准实践指南—人工智能伦理安全风险防范指引》, issued in June 2021, provides comprehensive guidance for organizations and individuals on preventing ethical and safety risks in AI activities, while the Code of Ethics for the New Generation Artificial Intelligence《新一代人工智能伦理规范》, issued in June 2021, integrates ethical principles into the AI lifecycle, promoting fairness, justice, and safety while avoiding bias and privacy issues.

China emphasises stakeholder responsibility in the ethical development of AI through the Opinions on Strengthening the Governance of Scientific and Technological Ethics 《关于加强科技伦理治理的意见》, issued in March 2022, which aims to enhance the governance of scientific and technological ethics and establishes mechanisms for ethical management. The Measures for the Review of Sci-tech Ethics (for Trial Implementation) 《科技伦理审查办法（试行）》, introduced in January 2023, further specifies the scope, procedures, and responsibilities for ethical reviews in scientific and technological activities, ensuring compliance and safety.

China's data protection regulatory regime consists of three key laws: the Cyber Security Law (CSL), the Data Security Law (DSL), and the Personal Information Protection Law (PIPL), along with various regulations and technical standards:

• The Cybersecurity Law《网络安全法》, issued in June 2016, mandates that network operators, including those providing generative AI services, comply with cybersecurity obligations to manage risks.
• The Data Security Law《数据安全法》, enacted in September 2021, regulates data processing activities, emphasizing data classification, risk assessment, and the protection of data security, which is crucial for AI technologies that rely on large datasets.
• The Personal Information Protection Law《个人信息保护法》, enacted in August 2021, regulates personal information processing activities, requiring that the collection of personal information should be limited to the minimum scope and the processing of personal information should have a clear and necessary purpose using methods that have the least impact on personal rights and interests.

While there are no specific regulations solely for AI data, it is addressed in various regulations. For example, the Regulation on Network Data Security Management 《网络数据安全管理条例》, introduced in January 2024, specifically requires that a network data handler providing generative AI services shall strengthen its security management of training data and training data handling activities and take effective measures to prevent and dispose of network data security risks.

In addition, the TC260 issued a series of national standards, providing specifications and guidance for the application of AI technology including:

1. Information security technology—General requirements for biometric information protection《信息安全技术 生物特征识别信息保护基本要求》
2. Information Security Technology - Security Requirements for Facial Recognition Data《信息安全技术 人脸识别数据安全要求》
3. Information Security Technology - Security Requirements for Automotive Data Processing《信息安全技术 汽车数据处理安全要求》
4. Cybersecurity Technology - Security Specifications for Data Annotation in Generative Artificial Intelligence《网络安全技术 生成式人工智能数据标注安全规范》
5. Cybersecurity Technology - Security Specifications for Pre-training and Optimisation Training Data in Generative Artificial Intelligence《网络安全技术 生成式人工智能预训练和优化训练数据安全规范》
6. Information Security Technology — Security Requirements of Voiceprint Recognition Data《信息安全技术 声纹识别数据安全要求》
7. Information Security Technology — Security Requirements of Gait Recognition Data《信息安全技术 步态识别数据安全要求》

China has established regulations to protect intellectual property (IP) and trade secrets, including certain basic laws as well as guidelines in the field of AI.

The Copyright Law《著作权法》, promulgated by the NPC Standing Committee in  September 1990 and last amended in November 2020, and the Regulations for Implementation of the Copyright Law《著作权法实施条例》,  promulgated by the State Council in August 2002 and last amended in January 2013, provide that Chinese citizens, legal persons, or other organisations enjoy copyright protection for their works, whether published or not, in fields which include works of literature, art, natural science, social science, engineering technology, and computer software. 

The Guidelines for Patent Applications Related to Artificial Intelligence (Trial) 《人工智能相关发明专利申请指引（试行）》, issued in December 2024 by the China National Intellectual Property Administration, provides detailed instructions for patent applications in the AI field. These guidelines cover types of patent applications, identification of inventors, standards for patentable subject matter, sufficient disclosure in the specification, considerations of inventiveness, and ethical issues. The guidelines emphasize that patent applications should reflect technical means to solve specific technical problems, inventors must be natural persons, and the specification should describe the technical solution in detail to enable implementation by skilled personnel.

The Anti-Unfair Competition Law of the People’s Republic of China《反不正当竞争法》, enacted in April 2019 by the National People’s Congress, defines “trade secrets” as technical and business information that is not known to the public, has commercial value, and is subject to confidentiality measures by the rights holder. Generative AI technologies, which involve data resources for training or learning, are closely related to trade secrets. Information related to large models during training, fine-tuning, and inference processes includes code, algorithms, and data.

On 10 February 2025, the Hangzhou Internet Court ruled in favor of Tsuburaya Productions, finding the defendant's AI platform liable for copyright infringement in allowing users to train and share models based on copyrighted material. The court ordered the platform to cease infringing activities and pay RMB 30,000 in damages.

China has implemented stringent regulations to control the export of AI technologies, ensuring their responsible use and preventing national security risks.

The Regulations on the Administration of Technology Import and Export 《技术进出口管理条例》, issued in November 2022 , and the Catalogue of Technologies Prohibited or Restricted from Export by China《中国禁止出口限制出口技术目录》, issued in December 2023, restricts certain AI related technologies on exports such as:

• Network early warning linkage response technology for large-scale network.
• Speech synthesis technology dedicated to Chinese and minority languages.
• AI interactive interface technology specialised in Chinese and minority languages.
• Intelligent examination paper scoring technology specially used for Chinese and minority languages.
• Personalised information push service technology based on data analysis.

The Regulations on the Export Control of Dual-use Items《两用物项出口管制条例》, issued in September 2024 by the State Council, serve as an important supporting rule for the Export Control Law. These regulations apply to the transit, transshipment, transportation, re-export, and export from special customs supervision areas or bonded areas of dual-use items.

The Export Control List of Dual-use Items of the People's Republic of China《两用物项出口管制清单》, issued in November 2024 by the Ministry of Commerce, includes items from various fields such as information technology, communication technology, AI, and quantum computing. The list specifically covers AI and automation technologies, quantum technologies, and semiconductor technologies.

China has established several key regulations to govern foreign direct investment (FDI), ensuring a stable and transparent environment for foreign investors:

• The Foreign Investment Law《外商投资法》, adopted in March 2019 by the National People’s Congress and effective from 1 January 2020, ensures that foreign investors receive treatment no less favourable than domestic investors, adhering to the principle of “national treatment” throughout the investment lifecycle, including in the AI sector.
• The Catalogue of Encouraged Industries for Foreign Investment (Edition 2022) 《鼓励外商投资产业目录（2022年版）》issued by the National Development and Reform Commission (NDRC) and the Ministry of Commerce (MOFCOM) in October 2022 and effective in January 2023, and The Special Administrative Measures (Negative List) for Foreign Investment Access (Edition 2024) issued in September 2024 by NDRC and MOFCOM and effective in November 2024《外商投资准入特别管理措施（负面清单）(2024年版)》, categorise industries into “encouraged”, “restricted”, and “prohibited". The Encouraging Catalogue and the Negative List are subject to review and update by the Chinese government from time to time. Under the current Encouraging Catalogue, certain AI technologies are listed as encouraged industries for foreign investment, such as manufacturing of AI-assisted medical equipment, wearable intelligent health equipment, medical imaging intelligent auxiliary diagnosis systems, manufacturing of smart consumer equipment such as wearable smart devices and smart unmanned aerial vehicles, research and development on the AI technologies including intelligent devices, robotics, neural network chips, and neuron sensors.
• The Measures for the Security Review of Foreign Investments《外商投资安全审查办法》, issued in December 2020 by the NDRC and the MOC, requires a security review on the foreign investors that affect or may affect national security, which include the investments in military industry, military industrial supporting and other fields relating to the security of national defence, and investment in important agricultural products, important energy and resources, important equipment manufacturing, important infrastructure, important transport services, important cultural products and services, important information technology and Internet products and services, important financial services, key technologies and other important fields relating to national security, and obtaining the actual controlling stake in the investee enterprise.