The field of artificial intelligence (“AI”) is at a tipping point. Governments and industries are under increasing pressure to forecast and guide the evolution of a technology that promises to transform our economies and societies. In this series, our lawyers and advisors provide an overview of the policy approaches and regulatory frameworks for AI in jurisdictions around the world. Given the rapid pace of technological and policy developments in this area, the articles in this series should be viewed as snapshots in time, reflecting the current policy environment and priorities in each jurisdiction.
On the sidelines of November’s APEC meetings in San Francisco, Presidents Joe Biden and Xi Jinping agreed that their nations should cooperate on the governance of artificial intelligence. Just weeks prior, President Xi unveiled China’s Global Artificial Intelligence Governance Initiative to world leaders, the nation’s bid to put its stamp on the global governance of AI. This announcement came a day after the Biden Administration revealed another round of restrictions on the export of advanced AI chips to China.
China is an AI superpower. Projections suggest that China’s AI market is on track to exceed US$14 billion this year, with ambitions to grow tenfold by 2030. Major Chinese tech companies have unveiled over twenty large language models (LLMs) to the public, and more than one hundred LLMs are fiercely competing in the market.
Understanding China’s capabilities and intentions in the realm of AI is crucial for policymakers in the U.S. and other countries to craft effective policies toward China, and for multinational companies to make informed business decisions. Irrespective of political differences, as an early mover in the realm of AI policy and regulation, China can serve as a repository of pioneering experiences for jurisdictions currently reflecting on their policy responses to this transformative technology.
This article aims to advance such understanding by outlining key features of China’s emerging approach toward AI.
I. China’s AI Vision & Approach
In China, the Communist Party (“CCP”) often sets the overall economic and policy direction that the government is expected to follow—a concept known as “top-level design,” or “顶层设计.” It is, therefore, notable that between March and July of 2023, two of three meetings of the highest political body in the CCP, the Politburo, resulted in directives setting the “safe” development of AI as a national priority.
Following the CCP’s direction, the Chinese government has sought to proactively foster domestic AI capabilities—which it sees as critical to the country’s economic development—while imposing robust regulations and technical standards to contain perceived risks. A key milestone in China’s AI policy framework, the 2017 New Generation AI Development Plan (“2017 AI Plan”) set a goal of making the country a global AI innovation center with an RMB 1 trillion industry by 2030. To achieve this goal, China’s 14th Five-Year Plan established a yearly 7% growth target for overall R&D expenditure and highlighted AI as one of seven frontier areas prioritized for technological breakthroughs. National and local governments have pursued industrial policies to enhance computing power and improve training datasets to further support the development of AI in recent years.
Meanwhile, the Chinese government has been addressing perceived AI-related political and social risks through requirements to file algorithms with the government (“算法备案”) and other security assessment and ethics review mandates. These measures are driven by the government’s intent to restrict access to information it deems sensitive or undesirable, protect against misinformation, and ensure adherence to data security and personal information protection laws.
Additionally, in line with China’s growing efforts to project power and shape a favorable external environment for its political and economic interests, President Xi and the Chinese leadership are actively promoting Chinese initiatives and approaches at the international level. China’s leaders recognize the global nature of AI technology and competition. At the third Belt and Road Forum in Beijing in October 2023, President Xi announced China’s proposal for a Global AI Governance Initiative. The accompanying communique called for equal opportunities for AI development irrespective of a country’s political system and the rejection of exclusive groups and ideological divides. Implicit in the initiative is a critique of U.S. efforts to limit China’s access to advanced technologies needed for its AI industry.
II. Major Domestic Policy & Regulatory Initiatives
In China, AI policy originates with the CCP and the State Council, which serves as the government’s principal administrative organ or cabinet. Multiple institutions are responsible for regulating and promoting AI under their leadership:
- Cyberspace Administration of China (“CAC”). CAC primarily controls online content and also participates in the development and enforcement of the country’s cybersecurity, data security, and privacy laws. CAC is a hybrid organization that spans both the Party and the state.
- Ministry of Science and Technology (“MOST”). MOST formulates strategies for China’s innovation-driven development. It released high-level ethics guidance related to AI and issues grants to support major AI R&D projects.
- Ministry of Industry and Information Technology (“MIIT”). MIIT is responsible for China’s industrial and information technology policies. It has endorsed important AI regulations and initiated multiple projects encouraging AI-related R&D.
- National Data Bureau (“NDB”). As part of a broader reorganization of government agencies in March 2023, the National People’s Congress approved the creation of a new National Data Bureau within the National Development and Reform Commission (“NDRC”). Although publicly available information is limited, NDB will likely play a critical role in China’s AI governance, as it will absorb certain responsibilities related to the management of national data infrastructure that were previously held by CAC.
A. Government Support for AI Development: Industrial Policies to Enhance AI Computing Power and Training Data
Under the 2017 AI Plan and the 14th Five-Year Plan, China has enacted a range of industrial policies at the national and local levels aimed at advancing the country’s AI capabilities, with a focus on enhancing computing power and building comprehensive training datasets.
Enhancing Computing Power for AI. In October 2023, MIIT and other agencies announced the Action Plan on the Development of High-Quality Computing Power Infrastructure. Among other things, the plan aims to increase China’s computing power to 300 EFLOPS by 2025. According to MIIT, China reached 197 EFLOPS, ranking second globally, as of 2023. The action plan includes the creation of 50 computing hubs by 2025 to boost advanced computing capabilities and improve data management, processing, and infrastructure. This initiative highlights the significant investment of the Chinese government in advanced computational infrastructure to meet future AI and other computing demands.
Following this national plan, local governments, such as Shanghai and Beijing, have introduced financial incentives and projects to enhance computing power. Shanghai, for instance, released a list of “new infrastructure” projects, many of which specifically target building computing power. A Beijing voucher program aims to lower the cost of AI computing power and support the use of AI in sectors like manufacturing, healthcare, finance, education, and transportation.
Building Comprehensive Training Data Sets. In 2021, China’s MIIT launched a Three-Year Action Plan for developing new data centers across the country. The plan focused on integrating various data sources, building and upgrading data centers, and creating a large-scale national data platform. The government encouraged local authorities to share data resources to advance AI, while protecting personal and business information security in the process.
Shanghai and Shenzhen are two cities that have taken significant steps in this direction. Shanghai has targeted the production of over 1,000 high-quality training datasets by 2025 for manufacturing, finance, shipping, and innovation-related industries. Shenzhen is enhancing AI data accessibility with a city-wide public data platform for high-quality Chinese language training datasets, and expanding the state-run Shenzhen Data Exchange that was launched in November 2022. These efforts demonstrate the Chinese government’s commitment to improving data access and quality for AI development.
B. Preventing AI-Related Social Risks through Robust Regulation: Security Assessment, Algorithm Filing, and Ethics Review Requirements
While promoting AI through industrial policies and incentives, the Chinese government is also seeking to contain the social and political risks of AI—particularly those related to the technology’s influence on public opinion and capacity for social mobilization—through regulation and technical standards. These efforts focus on three major areas: security assessments, algorithm filings, and ethics reviews.
Security Assessments and Algorithm Filings. In July, seven Chinese government agencies led by CAC released the Interim Administrative Measures for Generative Artificial Intelligence Services (“Interim GenAI Measures”), China’s first regulation specifically for generative AI. The measures require providers offering generative AI services to the public—specifically, those “with the attributes of public opinion or the capacity for social mobilization”—to comply with security assessments, algorithm filings, and other requirements that involve disclosing extensive information about their services. The Interim GenAI Measures also align with existing data-related laws like the Data Security Law and the Personal Information Protection Law.
Further details about security assessments are provided in the Draft Basic Security Requirements for Generative Artificial Intelligence Service (“Draft GenAI Security Requirements”), which were issued in October 2023 by the National Information Security Standardization Technical Committee (TC260)—a technical organization in China responsible for developing standards related to information security. This draft specifies that generative AI services offered to the public in China must undergo security assessments based on TC260 standards. The standards cover training data and model security, including statistical baselines for testing and specific criteria for refusing certain queries. Notably, service providers are prohibited from using unfiled “foundation models” for research and development. TC260 is in the process of drafting and gradually releasing a comprehensive set of technical standards to fortify China’s regulatory framework for AI security assessment.
Ethics Reviews. In October 2023, MOST and nine other agencies in China introduced the Scientific and Technological Ethics Review Regulation (Trial) (“Trial Ethics Regulation”), effective from December 1, 2023. This regulation, following earlier ethics guidance from the State Council, CCP, and MOST, focuses on the ethical review of activities involving personal information and challenges in life and health, ecology, public order, and sustainable development.
The regulation mandates that entities conducting AI research in “sensitive areas of technology” (currently undefined) form a scientific ethics review committee, with more detailed implementing rules to be formulated by relevant authorities, including local authorities. High-risk activities, such as R&D related to algorithms and models, applications, and systems that can influence public opinion or social awareness, require an additional expert review.
III. Looking Forward
The environment in China for AI is evolving rapidly. The Chinese government continues to release policies and regulations affecting the sector at a rapid pace, and the U.S. government continues to expand the use of trade controls targeting the development of AI in China. Going forward, industry players and policymakers may wish to keep an eye on developments in the following areas:
- Plans to draft a comprehensive Artificial Intelligence Law. China’s AI governance, traditionally iterative in nature, may see a major step in its evolution as China develops a comprehensive Artificial Intelligence Law, plans for which were announced in June 2023. The development of this law will involve multiple ministries, review by the State Council, and ratification by the National People’s Congress. While the timeline is uncertain, the law would aim to integrate existing regulations into a cohesive framework governing Chinese AI policy.
- U.S. restrictions and Chinese responses. China’s swift AI advancement is unfolding amid escalating U.S. export controls targeting the sector, particularly on advanced semiconductors. At the same time, the U.S. Congress is actively promoting American innovation in foundational technologies like AI, while also curbing the transfer of critical emerging technologies to “foreign entities of concern,” including China. Companies are advised to vigilantly monitor U.S. export controls and how China responds.
- Implementation of existing measures. China’s existing AI policies and regulations present significant compliance challenges for companies operating in China due to lack of guidance and precedent, overlapping and sometimes inconsistent requirements, and gaps in regulations. Companies should stay alert to revisions in the Draft GenAI Security Requirements and actively participate in related rule-setting and implementing processes to grasp essential information and pass AI service security assessments. With the unclear definition of “sensitive areas” in the soon-to-be effective Trial Ethics Regulation, companies should seek to engage in shaping local implementation rules. Additionally, as AI infrastructure policies evolve, companies should be alert to opportunities to take advantage of public-private partnerships and benefit from various incentive schemes.
IV. Thought Leadership
Our public policy and regulatory teams closely track and contribute to discussions around AI policy in China. Below is a selection of related articles from our public-facing blogs:
- Labeling of AI Generated Content: New Guidelines Released in China, August 30, 2023
- A Practitioner’s Guide to China’s New Generative AI Reg, August 18, 2023
- China Proposes Draft Measures to Regulate Generative AI, April 12, 2023
- China Reveals Plan to Establish a National Data Bureau, March 7, 2023
* * *