Regulation of AI in China
As Chinese tech companies emerged as worldwide important players in the IT industry, the legal regulation of the sector became a more and more pressing issue for the government as well. China has taken a number of steps over the last few years to develop data protection and IT security regulation, and finally, the topic of AI also got the attention of the state.
Regulation of AI in China
Photo: iStock
Vilmos Gábor Rádi 03/10/2023 09:58

As Chinese tech companies emerged as worldwide important players in the IT industry, the legal regulation of the sector became a more and more pressing issue for the government as well. China has taken a number of steps over the last few years to develop data protection and IT security regulation, and finally, the topic of AI also got the attention of the state.

The regulator’s first steps concentrated on specific areas, which quickly became important:

  1. Article 24 of the Chinese privacy legislation, the so-called Personal Information Protection Law of the People's Republic of China (PIPL)[1] paid attention to the automated decision making processes. It sets out ethical requirements, prohibits unjustified disparate treatment in business terms, and provides some specific protection for end users.
  2. The widespread usage of algorithmic recommendations generated the need for a separate legislation, which embodied as the Internet Information Service Algorithmic Recommendation Management Provisions[2], which requires such services to be ethical, moral, accountable and transparent. The document calls itself a recommendation, but it can be viewed more as a regulation, since the supervisory authority has the possibility to impose a fine based on the violation of the recommendation. The legal act requires companies to inform users if an algorithm plays a role in determining what information to display to them and to give users the opportunity to opt out of a targeted recommendation. In addition, the Regulation prohibits the use of algorithms that use personal data to offer different prices to different consumers for the same product/service.

Other actors in the organization of the state also felt the need to step up, so besides the above-mentioned obligatory acts, three recent recommendations/guidelines were issued:

  1. A "White Paper on Trustworthy Artificial Intelligence" issued by the Academy of Information and Communication Technologies[3], which touches the areas of development, support, practical usage of AI, as well as ethical and privacy implications.
  2. The Code of Ethics for the next generation of Artificial Intelligence, issued by the Ministry of Science and Technology[4], deals with issues of R&D, management, implementation, etc.
  3. The Guiding Opinions on Strengthening the Governance of Algorithms for Internet Services[5], which basically proposes to strengthen the management control systems of AI programs.

It is worth mentioning that the need for being capable of distributing internationally marketable AI software is so high, that the municipality of the special economy zone of Shenzen city made dire steps to ensure its primary status among other cities. Shenzen is home to several tech giants, which are all leaders in AI development, so this city decided not to wait for the state-level complex legislation; instead it issued its own local, municipality-level regulation[6] which ensures the support of R&D projects, and also calls for the development of local standards and certification mechanisms, the development of an ethical risk assessment and even creates a "Municipal Artificial Intelligence Ethics Committee".

As a quite recent event, the first obligatory regulation focusing on AI as a general phenomenon just came out on 11th April 2023, and its consultation period already ended on 10th of May.  The Administrative Measures for the Management of Generative Artificial Intelligence Services[7]. The legislation expects previous security assessments and content control of the software on the side of the service providers, anti-discrimination expectations in contrast to AI bias, requires intellectual property rights protection, and prevention of  unfair competition, spreading misinformation, physical and mental harm of other persons; also sets out rules about the quality of the training data, privacy, complaint process, and expects the service providers to identify the end users and also educate them


The author is a digital business lawyer and data protection expert, holds an LL.M


[1] Personal Information Protection Law of the People's Republic of China http://en.npc.gov.cn.cdurl.cn/2021-12/29/c_694559.htm [2023. 06. 19.]

[2] Internet Information Service Algorithmic Recommendation Management Provisions

http://www.gov.cn/zhengce/zhengceku/2022-01/04/content_5666429.htm [2023. 06. 19.]

[3] White Paper on Trustworthy Artificial Intelligence"  http://www.caict.ac.cn/english/research/whitepapers/202110/P020211014399666967457.pdf [2023. 06. 19.]

[4] The Code of Ethics for the next generation of Artificial Intelligence https://www.most.gov.cn/kjbgz/202109/t20210926_177063.html [2023. 06. 19.]

[5] Guiding Opinions on Strengthening Overall Governance of Internet Information Service Algorithms http://www.moe.gov.cn/jyb_xxgk/moe_1777/moe_1779/202109/t20210929_568182.html [2023. 06. 19.]

[6] Shenzhen Special Economic Zone Artificial Intelligence Industry Promotion Regulations https://law.pkulaw.com/chinalaw/eb370a7e0d9edd5e8ca8bb1a5fa6a5e7bdfb.html [2023.06.19.]

[7] Administrative Measures for the Management of Generative Artificial Intelligence Service http://www.cac.gov.cn/2023-04/11/c_1682854275475410.htm [2023. 06. 19.]

As we can see, China quickly developed its own legislation of AI; within a mere 2 years, from the first steps focusing on specific application areas, through guidelines, the state arrived to the first complex regulation of artificial intelligence.  The Cyberspace Administration of China plays a central role in the future processes, responsible for both issuing detailed rules and acting as supervisory authority.  The already issued obligatory rules and recommendations have similar main elements compared to the European draft regulations, like the requirement of risk assessments, quality of training data, privacy protection, etc., but also contains some Chinese specialties, like the need to identify the end user. As more and more machine learning softwares are issued by the Chinese tech companies, and huge amounts of yuans spent for development in each year, the Chinese legal regulation is required to both to ensure the AI software development is safe for the Chinese people and also transparent, ethical, and therefore marketable for the international market.

We use cookies on our website. If you consent to their use, we use them to measure and analyze the use of the website.
Information and Settings