China's Orchestrated Push to Build an AI Talent Superpower

China has constructed the most ambitious state-directed artificial intelligence talent pipeline in the world, moving from a single policy blueprint in 2017 to more than six hundred universities offering AI degrees by 2025, with technology giants operating joint laboratories inside campuses and digital human technologies beginning to reshape how students learn. The mobilization is without precedent in speed or scale. Within six years of designating AI a national strategic priority, the State Council, the Ministry of Education, and provincial governments cascaded a series of interlocking plans that restructured higher education curricula, funneled billions of yuan into university-industry research partnerships, and in 2025 began mandating AI literacy instruction for every primary and secondary school student in Beijing. The effort reflects a calculated bet that the country controlling the AI talent supply chain will dominate the industries of the coming decades. The architecture rests on foundational policy documents, a massive expansion of university-level AI programs, deep entanglement between technology corporations and the academy, and an emerging layer of digital human applications in education that ranges from avatar-based lecturers to AI-powered museum guides. Each element reinforces the others, creating a feedback loop in which policy drives institutional capacity, institutional capacity attracts corporate investment, and corporate technology reshapes pedagogy.

The starting gun was the New Generation Artificial Intelligence Development Plan (新一代人工智能发展规划), issued by the State Council on July 8, 2017. The plan declared AI the defining strategic technology of international competition and laid out a three-step roadmap: keep pace with leading AI nations by 2020, achieve major breakthroughs by 2025, and become the world's primary AI innovation center by 2030. For the talent pipeline specifically, it mandated the establishment of AI-related courses in primary and secondary schools, the gradual promotion of programming education, the expansion of AI graduate enrollment, and the recruitment of global high-end talent to form what it called a "talent highland." It also called for training "vertical compound talents" who master AI theory through to product development, and "horizontal compound talents" who can apply AI across economics, governance, law, and standards-setting. The Ministry of Education translated those directives into operational terms less than a year later with the Higher Education Artificial Intelligence Innovation Action Plan, released on April 2, 2018, which instructed universities to build AI as a first-level discipline, develop an "AI+X" interdisciplinary training model spanning computer science, mathematics, biology, psychology, and law, and produce textbooks and massive open online courses in machine learning, neural networks, computer vision, natural language processing, and knowledge engineering. The plan designated smart education as a priority application domain and set milestones for 2020, 2025, and 2030, with the ultimate goal that universities would become core forces in building major AI innovation centers.

In April 2025, Xi Jinping underscored the urgency at the twentieth Politburo collective study session, calling for AI education across all school stages and society-wide general education. The State Council subsequently issued the Opinions on Deeply Implementing the "AI+" Action, demanding what it called "super-conventional construction of new models for cultivating leading talent." Beijing's municipal education commission operationalized these directives in its 2025 to 2027 AI Education Work Plan, which mandates a minimum of eight hours per year of AI instruction for every student from elementary through high school beginning in autumn 2025, the training of one hundred expert AI teachers and one thousand backbone instructors, and the use of virtual digital human technology combined with large models to create high-level AI teachers.

Before 2019, the closest undergraduate major to artificial intelligence was Intelligent Science and Technology (智能科学与技术), first offered at Peking University (北京大学) in 2003. The formal AI undergraduate major did not exist until the Ministry of Education approved the first batch of thirty-five universities to offer it in 2019. Growth was explosive: 180 additional universities gained approval the following year, more than three hundred had the major by early 2021, and the cumulative total reached 502 universities by the end of 2023. By 2025, more than 626 institutions offered AI-related degrees, and twenty-four new AI-adjacent specialties were added to the national catalogue in that year alone. The leading research universities did not wait for the formal major designation. Tsinghua University (清华大学) established its AI Research Institute in June 2018 under the directorship of Academician Zhang Bo, with Turing Award laureate Yao Qizhi chairing its academic committee. In April 2024, Tsinghua elevated its commitment by founding the College of AI with Yao Qizhi as dean, organized around two pillars: "AI Core," covering foundational theory, embodied intelligence, multimodal perception, and future computing paradigms including photonic and quantum systems, and "AI+," targeting safety governance, public health, smart manufacturing, and energy systems. The university declared its ambition to rival MIT and Stanford within five years and announced plans to recruit fifty full-time professors.

Peking University launched its AI Research Institute in April 2019 and followed with the School of Intelligence (智能学院) in December 2021, consolidating its Intelligent Science Department, the Wang Xuan Computer Institute, and the AI Research Institute under the leadership of the globally recognized computer vision scholar Zhu Songchun. The school pursues general artificial intelligence as its long-term objective and runs two elite experimental classes, the "Tong Class" for general AI and the "Zhi Class" for intelligent science. Zhejiang University (浙江大学), which had been training AI graduate students since 1978, inaugurated its AI School in August 2024 and will operate it as an independent secondary college from July 2025, with research organized around intelligent theory and systems, intelligent design, and autonomous systems. Renmin University of China (中国人民大学) took a distinctive humanities-inflected approach by establishing the Gaoling School of Artificial Intelligence (高瓴人工智能学院) in April 2019, funded by a 300 million yuan donation from Hillhouse Capital founder Zhang Lei. The school focuses on integrating AI with law, economics, and media, and has ranked first globally in internet and information retrieval on CSRankings for three consecutive years from 2022 through 2024.

Supporting this university ecosystem is the Beijing Academy of Artificial Intelligence (BAAI, 北京智源人工智能研究院), a non-profit research institution founded on November 14, 2018, under the joint guidance of the Ministry of Science and Technology and the Beijing Municipal Government. Its founding members include Peking University, Tsinghua University, the Chinese Academy of Sciences, Baidu (百度), and ByteDance (字节跳动). BAAI operates the Zhiyuan Scholars program for leading scientists, hosts an annual international conference, and has produced notable research outputs including the WuDao 2.0 large model with 1.75 trillion parameters in 2021, the FlagAI open-source toolkit for large model training, and BAAIWorm, a computational model of the C. elegans nervous system published as a cover story in Nature Computational Science in December 2024. The United States placed BAAI on its Entity List in March 2025.

China's seven defense-oriented universities, collectively known as the "Seven Sons of National Defense" (国防七子), also figure prominently in the AI talent landscape. All subordinate to the Ministry of Industry and Information Technology, they are Beihang University (北京航空航天大学), Beijing Institute of Technology (北京理工大学), Harbin Institute of Technology (哈尔滨工业大学), Harbin Engineering University (哈尔滨工程大学), Nanjing University of Aeronautics and Astronautics (南京航空航天大学), Nanjing University of Science and Technology (南京理工大学), and Northwestern Polytechnical University (西北工业大学). These institutions supply an outsized share of graduates to defense-related state enterprises and devote substantial research budgets to military applications. They formalized their cooperation in 2017 through the G7 Alliance, all seven have been placed on the U.S. Entity List, and at least three had established AI institutes before the 2017 national plan was published.

Talent recruitment from overseas has historically been supported by two complementary programs. The Thousand Talents Plan (千人计划), launched in December 2008 by the CPC Central Committee Organization Department, aimed to recruit approximately two thousand high-level overseas experts over five to ten years across all strategic technology fields, including AI. Following heightened scrutiny from U.S. law enforcement agencies, the program was effectively suppressed in 2018: a directive ordered institutions not to publicly use the name, Chinese search engines and social media platforms blocked the term by 2020, and the program's functions were absorbed into successor initiatives. The separate Plan 111 (高等学校学科创新引智计划), launched in 2005 by the Ministry of Education and the State Administration of Foreign Experts Affairs, takes a different approach, recruiting roughly one thousand overseas academic leaders to build approximately one hundred world-class discipline innovation bases at Chinese universities, with each base receiving at least 1.8 million yuan annually over five-year cycles. More than two hundred bases had been established by 2017. Though neither program is AI-specific, both have been instrumental in populating university AI departments with internationally trained researchers.

The entanglement between China's technology corporations and its universities goes far beyond sponsorship. It constitutes a structural integration that Chinese policymakers describe using the "triple helix" model of university-industry-government collaboration, and the most vivid institutional expression of this model is Zhejiang Lab (之江实验室). Founded in 2017 as a joint venture of the Zhejiang Provincial Government, Zhejiang University, and Alibaba Group (阿里巴巴集团), funded with 30 billion yuan from government sources and occupying a dedicated 247-acre campus, Zhejiang Lab conducts research in intelligent perception, computation, networking, and systems. It stands as perhaps the clearest example in the world of a purpose-built triple-helix AI institution. Alibaba's university partnerships extend well beyond Zhejiang Lab. The Alibaba-Zhejiang University Joint Institute of Frontier Technologies, established in early 2017, operates sub-laboratories in cybersecurity, computer vision, the Internet of Things, data mining, database technology, and knowledge graphs, and has won the Zhejiang Provincial Technology Invention Award at the first-prize level. In 2022, Alibaba added an AI safety joint lab with Zhejiang University focused on large language model security and multilingual safety evaluation. The company has also built joint laboratories with Tsinghua University on natural human-computer interaction, with Peking University on AGI and digital humans through the Alimama AI Innovation Joint Lab, and with the University of Science and Technology of China on cognitive AI and privacy-preserving computing. Since 2022, Alibaba's DAMO Academy (达摩院) has collaborated with more than eighty universities on approximately three hundred research projects, producing around five hundred papers at top conferences.

Tencent (腾讯) built its first university-enterprise AI joint laboratory with the Chinese University of Hong Kong, Shenzhen in June 2018, co-directed by Academician Luo Zhiquan and Tencent AI Lab Director Zhang Tong, with provisions for jointly funded doctoral students. Tencent's Rhino-Bird programs sponsor twenty to thirty university research projects annually, fund fifteen or more visiting scholars per year, and support the CCF-Tencent Rhino-Bird Fund for young scholars, sustaining partnerships with more than twenty Chinese universities. Tencent's YouTu Lab, focused on computer vision, co-founded the Global AI Academic Alliance with institutions including MIT, Tsinghua, and Shanghai Jiao Tong University. Baidu (百度) anchors the government end of the triple helix most directly. In February 2017, the National Development and Reform Commission approved the National Engineering Laboratory of Deep Learning Technology and Application, led by Baidu and co-built with Tsinghua University, Beihang University, the China Electronics Standardization Institute, and the China Academy of Information and Communications Technology, making it China's first national-level AI engineering laboratory. Through its PaddlePaddle (飞桨) open-source deep learning platform, Baidu has partnered with more than five hundred universities, trained over two thousand instructors, and helped 226 institutions establish credit-bearing AI courses. SenseTime (商汤科技), which spun out of the Chinese University of Hong Kong's multimedia lab in 2014, co-founded the Global AI Academic Alliance in 2018 and was designated China's National Open Innovation Platform for Next-Generation AI on Intelligent Vision.

DeepSeek (深度求索), founded in Hangzhou in July 2023 as a spinoff from the High-Flyer Capital hedge fund's AGI research laboratory, occupies a distinctive position in China's AI ecosystem that requires precise characterization. DeepSeek is a large language model company, not an avatar or digital human platform. Its releases include DeepSeek-V3, a 671-billion-parameter model trained for approximately 5.6 million dollars, and DeepSeek-R1, a reasoning model released in January 2025 under an MIT license. Multiple universities have incorporated DeepSeek into their operations: Shenzhen University launched an AI general course built on DeepSeek in partnership with Tencent Cloud, Zhejiang University opened a public course series, Shanghai Jiao Tong University upgraded its AI learning tools with the model, and Renmin University applied it across teaching, research, and campus administration. These are uses of DeepSeek as a general-purpose AI tool, however, not as infrastructure for digital classrooms or automated lecture systems.

The integration of digital human technology into Chinese education remains in its early stages but is advancing rapidly, driven by policy mandates, falling production costs, and improvements in large language model capabilities. The most rigorously documented large-scale deployment is at Zhejiang University, where the School of Marxism developed the "AI Ideological Course One-Click" (AI思政课一点通) system for the 2024 to 2025 autumn-winter semester. Built with technology from Chaoxing (超星), the system uses digital human lecturers modeled on actual course instructors to deliver lectures on topics including generative AI ethics and legal literacy. The system served more than five thousand first-year undergraduates and has been described as the first large-scale domain-specific AI model deployed in China's ideological and political education curriculum. A separate initiative at the same university saw Professor Chen Wenzhi create a digital clone of himself as an AI teaching assistant for computer architecture courses, built from recordings of past lectures. At Zhejiang University of Finance and Economics (浙江财经大学), Professor Yao Jianrong, who retired in October 2024 after nearly forty years of teaching, collaborated with the university and enterprise partners to create a highly realistic digital avatar of himself. Recorded in September 2024, the digital human can deliver lectures from text input and convert content into multiple languages including English and French. Yao also connected the system to a large language model for real-time question-and-answer sessions with students.

In the medical field, Fudan University's Zhongshan Hospital unveiled a digital avatar of Academician Ge Junbo (葛均波), one of China's most prominent cardiologists, at the nineteenth Oriental Congress of Cardiology on May 30, 2025. The avatar, powered by the CardioMind large model (观心大模型), a cardiovascular-specific AI system enriched with clinical guidelines and trial data, can conduct patient consultations and speaks with Ge's characteristic Shandong accent. During its demonstration, the digital physician correctly diagnosed a case of hereditary hemorrhagic telangiectasia type II, a rare condition. Regional initiatives are channeling digital human technology toward economic development and cultural preservation as well. The Yangtze River Delta (Yancheng) Digital Audiovisual Industry Base (长三角（盐城）数字视听产业基地), located in the Yannan High-tech Zone of Yancheng, Jiangsu, opened formally on September 6, 2023, with a total planned investment of one billion yuan. By late 2025, it housed 121 enterprises employing more than five thousand professionals, with tenants including Baidu, NetEase (网易), Tencent, and SenseTime. The base contains Baidu's first digital human industry base nationwide, the Baidu Intelligent Cloud VR Industry Empowerment Center, co-built with local authorities. It partners with local institutions including Yancheng Normal University on order-based talent cultivation programs for digital human and metaverse skills. In Wuhan, the AI startup Shuming Technology (数命科技), founded by Silicon Valley returnee Robert Luo, developed an AI museum guide digital human named Jing Xiaochu (荆小楚), a character designed in traditional Chinese aesthetic style. The system is embedded with knowledge of 110,000 cultural relics, can communicate in 143 languages, and provides route planning and artifact explanations. It has been deployed in more than thirty museums across China. In Nanjing, Beijingdonglu Primary School launched a digital teaching assistant named Xiao Bei (小北) in March 2023 as part of the city's twenty-six-school AI education pilot, capable of answering student questions across subjects and providing immersive scenario-based teaching.

The broader digital human market in China is growing at a pace that suggests education-sector adoption will accelerate. IDC valued the Chinese AI digital human market at 41.2 billion yuan in 2024, an 85.3 percent year-over-year increase, and projects it will reach 250.5 billion yuan by 2029. China's regulatory apparatus has moved in parallel with technology deployment rather than lagging behind it. The Provisions on the Administration of Deep Synthesis Internet Information Services, jointly issued by the Cyberspace Administration of China (CAC, 国家互联网信息办公室), the Ministry of Industry and Information Technology, and the Ministry of Public Security on November 25, 2022, took effect on January 10, 2023. The regulation defines "deep synthesis" broadly to encompass the use of deep learning, virtual reality, and other generative or synthetic algorithms to create text, images, audio, video, and virtual scenes. It mandates real identity verification of users, content review of inputs and outputs, mandatory labeling and watermarking of all synthesized content, and explicit consent before editing biometric information such as faces and voices. The deliberate choice of the term "deep synthesis" (深度合成) rather than the narrower "deepfake" reflected an intentional expansion of regulatory scope to cover all generative AI content.

Subsequent regulations have tightened the framework. The Interim Measures for the Management of Generative AI Services took effect on August 15, 2023, establishing a registration system through which 748 generative AI services had completed CAC filing by the end of 2025. In December 2025, the CAC published for public comment the Provisional Measures on the Administration of Human-like Interactive Artificial Intelligence Services, which targets AI services that simulate human personality traits and communication styles for emotional interaction. The draft requires clear disclosure that users are interacting with AI, mandates a two-hour continuous use warning, requires systems to detect user dependency and extreme emotional states, and stipulates that human staff must intervene when suicide or self-harm risk is detected. The Ministry of Education has issued its own guidance specific to AI in the classroom, publishing the Guide for Generative AI Use in Primary and Secondary Schools in 2025, which explicitly endorses using generative AI-based digital human teachers to provide teaching support in teacher-shortage areas. The ministry selected 184 schools as AI education pilot bases in December 2024, and by 2025, twenty-three provincial education departments had deployed AI education initiatives.

China's AI talent pipeline is not a single program but a layered system in which national strategic plans, university restructuring, corporate laboratory embedding, regional industrial bases, and regulatory frameworks operate as interlocking components. The speed of the build-out, from zero formal AI undergraduate majors in 2018 to more than six hundred institutions offering them by 2025, reflects both the advantages and the risks of centralized mobilization. The integration of digital human technologies into education, while still largely experimental, is moving from isolated demonstrations toward systemic deployment, pushed by explicit policy language in documents like Beijing's 2025 to 2027 plan that instructs schools to use virtual digital humans powered by large models as AI teachers. The most significant development may be the least visible: the structural fusion of corporate AI research with university operations, exemplified by arrangements like Zhejiang Lab and the National Engineering Laboratory of Deep Learning, which blur the boundaries between public research, private innovation, and state direction in ways that have no direct parallel in Western higher education systems.

[Apr 2026]