Hipolito Calero was an inaugural research intern at CSET during the summer of 2024 and is a senior at Florida International University in Miami, Florida.
Introduction
Policymakers worldwide are grappling with how to govern the development and use of artificial intelligence. China is an especially important country with regards to the development and governance of AI systems. Beijing has described AI as a “major strategic opportunity” for China and has called for the country to lead the world in AI by 2030. China has also been one of the first countries globally to introduce regulations that govern the use of AI systems, including some of the earliest and most detailed regulations governing recommendation algorithms put into effect in 2021.
This blog post provides analysis of the key actors involved in the governance of AI systems as specified in five major Chinese AI governance proposals. The analysis focuses on these key actors and responsibilities because they are pivotal to realizing China’s AI governance objectives. By analyzing the actors involved in China’s AI governance proposals, we seek to contribute to a better understanding of China’s changing approach to AI governance. We find that while older proposals tend to lack clarity and specificity with regards to AI governance actors, more recent proposals better specify AI governance actors and assign them detailed responsibilities.
Methodology
We conducted a literature review of Chinese AI governance proposals, drawing from CSET translations and previous analysis of China’s AI governance landscape conducted by the Carnegie Endowment for International Peace.1 From our review, we chose five AI governance proposals to further analyze. These proposals do not represent a comprehensive set of Chinese AI governance proposals. However, they cover a wide range of AI governance issue areas and sectors, are authored by different organizations, and were published between 2016 and 2024. These characteristics help us better understand how Chinese entities have delegated responsibilities to various actors in the Chinese AI governance ecosystem at different points in time. Table 1 displays the selected proposals in chronological order and includes a brief summary of each.
Table 1: Selected Proposals
| Title | Author(s) | Date of Release | Summary |
| “Internet+” Artificial Intelligence Three-Year Action and Implementation Plan (Internet+ Plan) | National Development and Reform Commission (NDRC); Ministry of Science and Technology (MOST); Ministry of Industry and Information Technology (MIIT); and Cyberspace Administration of China (CAC) | May 2016 | Describes a national strategy for the early development of China’s AI industry over three years, from 2016 to 2018. The plan encourages the growth of AI development and technologies in China. |
| New Generation AI Development Plan (Generation AI Development Plan) | State Council | July 2017 | Outlines a timeline and strategy to advance AI development and build out AI regulations in China. The plan pledges that China will become the top player in AI development by 2030. |
| “Trustworthy AI” White Paper | China Academy of Information and Communications Technology (CAICT) and JD Explore Academy | July 2021 | Focuses on defining the trustworthiness of an AI system. The paper analyzes potential paths that concerned parties could take to test and achieve an AI system that is reliable, transparent, and controllable. |
| Interim Measures for the Management of Generative Artificial Intelligence Services (Measures for Generative AI Services) | Cyberspace Administration of China (CAC) | July 2023 | A regulation that was promulgated and came into effect on August 15, 2023. Establishes measures to apply to most forms of generative AI services and applications, as well as the obligations of AI providers when building AI services and products. |
| Artificial Intelligence Law of the People’s Republic of China (Draft for Suggestions from Scholars) | Drafting Expert Group comprised of Professor Zhang Linghan of China University of Political Science and Law; Professor Yang Jianjun of Northwest University of Political Science and Law; Senior Engineer Cheng Ying of ChinaAcademy of Information and Communications Technology; Associate Professor Zhao Jingwu of Beijing University of Aeronautics and Astronautics; Associate Professor Han Xuzhi of East China University of Political Science and Law; Professor Zheng Zhifeng of Southwest University of Political Science & Law; and Associate Professor Xu Xiaoben of Zhongnan University of Economics and Law | March 2024 | Draft law that aims to promote innovation in AI technology, develop a healthy AI industry, and regulate AI products and services. Contains provisions for managing AI systems, safeguarding national security and public interest, and protecting the rights and interests of individuals and organizations. |
Key Findings
Older proposals do not specify actors involved in AI governance.
Early proposals, including the “Internet+ Plan” and the “New Generation AI Development Plan,” mention a number of ambitious undertakings but fail to specify the actors responsible for these projects. For example, the “Internet+ Plan,” published in 2016, highlights several ambitious key projects, including building platforms for AI resources and innovation, intelligent product innovations (such as intelligent vehicles, smart homes, and unmanned systems), world-leading AI enterprises, a standardization system, and a safe AI industry ecosystem. This particular proposal does not specify who is directly responsible for carrying out the projects and only briefly references academia’s role in promoting Chinese AI development.
A year after the release of the “Internet+ Plan,” China’s State Council released the “New Generation AI Development Plan.” Like the “Internet+ Plan,” the “New Generation AI Development Plan” sets ambitious goals for China’s AI sector. These objectives include developing AI industries; promoting intelligent product innovations; increasing AI investments, research, and projects; and encouraging talent training for AI knowledge and applications. However, the proposal remains vague about the actors who will be responsible for achieving these objectives.
Recent proposals designate responsibilities to specific actors.
In contrast to the proposals above, the “Trustworthy AI White Paper,” “Measures for Generative AI Services,” and the “Artificial Intelligence Law of the People’s Republic of China”—each published after 2021—more clearly assign responsibilities to specific actors.
The “Trustworthy AI White Paper,” for example, groups actors into different categories which include “Corporate,” “Industry,” “Government,” and “Technology Research.” These actors include employees and managers in the “Corporate” category, as well as third-party organizations and insurance institutions, which can share the responsibility of managing risk and provide compensation for losses from AI accidents, in the “Industry” category.
A substantial portion of the “Trustworthy AI White Paper” is devoted to assigning responsibilities to actors that fall into the “Corporate” and “Industry” categories. The “Corporate” category of AI governance actors refers to enterprises that develop AI technologies, products, and services. Responsibilities assigned to “Corporate” actors relate to incorporating trustworthy AI into the corporate culture. The “Industry” category of AI governance actors includes parties that manage the system of AI standards, evaluation, and verification as well as cooperation and exchange. Actors in the “Industry” category are responsible for evaluating and verifying AI services and products. The “Trustworthy AI White Paper” is the first of the five proposals analyzed to assign AI governance responsibilities to a defined set of actors.
“Measures for Generative AI Services” also delegates responsibilities to specific actors but assigns AI governance tasks to government bodies rather than corporate or industry players. Actors in this proposal include the Departments of Internet Information; Reform and Development; Education; Science and Technology; Industry and Informatization; Public Security; Radio and Television; and Press and Publication. Responsibilities allocated to departments include strengthening the management of generative AI services, collaborating and communicating with AI service providers, and developing new AI regulations. “Measures for Generative AI Services” is also the first proposal to assign responsibilities to AI providers, which are defined as organizations and individuals (such as researchers and developers) that provide generative AI services. The obligations assigned to AI providers include tagging images and videos produced by generative AI, optimizing training data, and counteracting illegal content.The highlighting of key departments and AI providers that play a role in AI governance is a departure from previous proposals that left the actors responsible for AI governance activities unspecified.
One of the most recent proposals, the “Artificial Intelligence Law of the People’s Republic of China,” references several actors, including AI developers, researchers, and users; companies that use AI systems; and “the State” and its organs. The responsibilities delegated to these actors are the most detailed out of all five proposals. For example, AI developers, providers, and users are tasked with assessing and explaining the safety risks of AI products and services. Companies using AI systems are prohibited from using the technology to infringe on the rights of workers. This proposal represents a shift toward more clearly delegating responsibilities to a wide range of AI governance actors.
At the same time, while recent proposals tend to include more specificity than older proposals, ambiguity remains. The “Artificial Intelligence Law of the People’s Republic of China,” for example, includes responsibilities for “the State”—a term presumably referring to China’s party-state apparatus—in many of the law’s provisions, but does not designate those responsibilities to specific state organs. Many of the provisions direct “the State” to establish new mechanisms, departments, or initiatives to assist with AI governance. In other words, China’s approach to AI governance is dynamic and will continue to change.
Outlook
This blog post provides a brief introduction to actors who play a role in China’s AI governance ecosystem. While not a comprehensive analysis of China’s AI governance landscape, this blog post can be used as a stepping stone for further research. Our analysis of the five Chinese AI governance proposals finds that newer proposals specify actors and delegate responsibilities to these actors, whereas older proposals do not specify actors involved in AI governance. Analysts can use these findings to further investigate key actors and their responsibilities in China’s AI governance ecosystem. As policymakers around the world decide who should bear responsibility for harnessing the potential of AI while minimizing its risks, they may benefit from a comparative analysis of the actors and responsibilities mentioned in different countries’ AI governance proposals. This blog post could be a first step toward comparing actors and responsibilities in Chinese AI governance proposals to those in policy documents from other countries. Policymakers should consider the instructive value of other nations’ governance strategies when fashioning an AI governance model.