China’s Digital Humans and the Values That Shape Them

China ranks as the world's second most powerful nation in artificial intelligence, trailing only the United States but leading in several important metrics. As of 2023, China accounted for 69.7 percent of all global AI patent grants, and Chinese researchers produced 23.2 percent of global AI publications and 22.6 percent of all AI citations, ahead of both Europe and the United States in volume. In 2024, while the United States produced 40 notable AI models to China's 15, Chinese models reached near-parity with their American counterparts on major performance benchmarks, and nine of the top ten open-weight models globally originated from China. China also installed more than half the world's industrial robots. Where the gap remains widest is private investment, with the United States attracting $109.1 billion in 2024 compared to China's $9.3 billion in private funding alone, though total Chinese AI investment across all sources reached an estimated $84 to $125 billion in 2025.

The Chinese Communist Party has established a direct and explicit connection between its official value system and the governance of technology. The Core Socialist Values, formally proposed at the 18th National Congress of the CPC in November 2012, comprise twelve values organized across three levels: prosperity, democracy, civility, and harmony at the national level; freedom, equality, justice, and rule of law at the social level; and patriotism, dedication, integrity, and friendliness at the individual level. These values have been systematically extended into technology regulation. The Interim Measures for Generative AI Services, effective August 15, 2023, mandate in Article 4 that providers uphold the Core Socialist Values. The draft Measures for the Management of Digital Virtual Human Information Services, released by the Cyberspace Administration of China on April 3, 2026, contain the same mandate in their own Article 4. The December 2025 draft on human-like interactive AI services requires that training datasets be consistent with socialist core values and reflect excellent traditional Chinese culture. Xi Jinping has personally reinforced this linkage on multiple occasions, stating at a November 2025 Politburo collective study session on cyberspace governance that cyberspace must become an important position for thought guidance, moral cultivation, and cultural inheritance.

Academic analysts have identified a strong Confucian dimension in China's approach to AI governance, though the government itself does not typically describe its framework in those terms. Official documents reference "excellent traditional Chinese culture" rather than Confucianism by name. Nevertheless, the concept of harmony, deeply rooted in Confucian thought, is embedded throughout China's AI governance architecture. The Ministry of Science and Technology's 2019 Governance Principles for New Generation AI listed "harmony and friendliness" as its first principle. The Beijing AI Principles of the same year included "harmony and cooperation," and the 2021 Ethical Norms for New Generation AI promoted "harmony and safety." Scholars have observed that China's AI governance model prioritizes collective welfare and government-led stewardship in ways that reflect Confucian ideas of humaneness, appropriateness, and hierarchical responsibility.

The State Council's New Generation Artificial Intelligence Development Plan, released on July 20, 2017, set out a three-phase strategy for China to become the world's primary AI innovation center by 2030, with core AI industry targets rising from 150 billion yuan by 2020 to one trillion yuan by 2030 and related industries reaching ten trillion yuan. This plan has been supplemented and extended by subsequent policies, including the 14th Five-Year Plan's target for digital economy core industries to reach ten percent of GDP by 2025, and an August 2025 State Council initiative setting goals of 70 percent AI agent penetration by 2027 and exceeding 90 percent by 2030. A $47.5 billion semiconductor fund was launched in 2024 to support AI chip independence. The private sector has responded at scale. Huawei spent $24.77 billion on research and development in fiscal year 2024, representing 20.8 percent of revenue, with 113,000 R&D employees comprising more than half its workforce. Alibaba pledged approximately $53 billion on AI and cloud infrastructure over three years beginning in 2025. Tencent invested 70.69 billion yuan in R&D and increased its capital expenditure by 221 percent year over year, integrating its Hunyuan large language model into more than 700 internal applications. ByteDance committed roughly $21 billion to AI infrastructure in 2025, with plans to increase to $23 billion in 2026.

China's regulatory framework for digital humans has developed through a layered sequence of increasingly specific measures. The Deep Synthesis Provisions, promulgated on November 25, 2022 and effective January 10, 2023, were jointly issued by the CAC, the Ministry of Industry and Information Technology, and the Ministry of Public Security. These provisions defined deep synthesis technology to include digital characters and virtual scenes, required AI-generated content labeling, and mandated consent for editing biometric information. The Generative AI Interim Measures followed in July 2023, issued by seven ministries, covering text, image, audio, and video generation and requiring algorithm filing with the CAC. By December 2025, 748 generative AI services had completed this filing process. The April 2026 draft regulation on digital virtual humans represents the first standalone regulatory instrument dedicated to the category. Its 27 articles address biometric consent requirements, special protections for minors under 14, a prohibition on virtual intimate relationships for those under 18, mandatory continuous display of a "digital human" label, and standard content prohibitions covering national security, separatism, terrorism, ethnic discrimination, violence, historical nihilism, and false information. The official definition contained in Article 25 specifies that digital virtual humans are virtual digital images that simulate human appearance using technologies such as computer graphics, digital image processing, or artificial intelligence, possessing characteristics such as voice, behavior, interactive ability, or personality. Pure text-only chatbots do not qualify under this definition, a distinction reinforced by the existence of the separate December 2025 draft covering human-like interactive AI that operates through text or voice without visual representation. National standards remain under development, with no final national standard specifically defining digital humans yet published.

The regulations governing digital human content impose legal obligations on service providers rather than constituting design specifications for the digital humans themselves. The April 2026 draft requires that the provision and use of digital virtual human services uphold the Core Socialist Values, comply with laws, preserve national security and the societal public interest, and conform to social mores, ethics, and morals. Content prohibitions focus on specific categories of harmful material. This regulatory architecture shapes the operating environment within which digital humans function but does so through provider compliance obligations rather than through mandated technical programming of the entities themselves.

Digital humans have found concrete application in Chinese healthcare. Tsinghua University's Agent Hospital project, with research published in May 2024 and an inauguration ceremony in April 2025, deployed 14 AI doctor agents and four AI nursing agents in integration with Chang Gung Hospital, later expanding to 42 AI doctors across 21 specialties through a spin-out company called Zijing Zhikang. The system achieved 93.06 percent accuracy on MedQA benchmarks. Ping An Health launched its "Ping An Xin Yi" platform on February 26, 2025, creating digital avatars of real-world doctors trained on databases covering 37,000 diseases and 420,000 terms, achieving 95.1 percent accuracy in AI-aided diagnosis and accumulating more than 35 million family doctor memberships by mid-2025. Gushengtang, a traditional Chinese medicine provider with 80 clinics in China and Singapore, deployed a digital avatar of the practitioner Li Hao covering eight specialties and reducing decision-making time to approximately six minutes per consultation. By February 2025, the DeepSeek-R1 model had been deployed in more than 260 hospitals across 93.5 percent of China's provinces, and the National Health Commission mandated that all tertiary hospitals integrate AI-assisted diagnosis by 2025.

In education, digital human avatars represent an emerging application within a broader AI-in-education movement. A digital Confucius interacts with fourth-grade students at Xinyilu Primary School in Qinhuangdao, Hebei Province, where every classroom is equipped with smart interactive boards and AI virtual assistants. iFLYTEK's "Curiosity Window" platform enables students to engage in natural-language dialogue with virtual personas of renowned scientists and cultural figures, while the company's SPARK Smart Blackboard, deployed in more than 50,000 schools and serving over 130 million teachers and students, features scenario-based virtual avatars. The National University of Defense Technology developed a Digital Teacher System creating one-to-one simulations of a teacher's personal image combined with a question-and-answer model. In rural Guizhou Province, the Hongyan AI Project from Beijing University of Posts and Telecommunications provides personalized digital mentoring for students in Changshun County using Socratic interaction logic. The Ministry of Education selected 184 schools as AI education pilot bases in February 2024 and, together with eight other departments, released guidelines in April 2025 on accelerating education digitalization, with a target of full AI coverage in primary and secondary education by 2030. As of that date, China had provided AI literacy training to 2.97 million teachers and developed more than 700 specialized AI teaching tools.

Digital humans have appeared at high-profile international events, though their deployment has been primarily commercial and cultural rather than diplomatic. Xinhua News Agency and Sogou launched the world's first AI news anchors at the fifth World Internet Conference in Wuzhen on November 7, 2018, with a female AI anchor, Xin Xiaomeng, following in February 2019. During the Beijing 2022 Winter Olympics, approximately 20 digital humans debuted, including a Baidu-powered AI sign language guide, the Alibaba DAMO Academy virtual influencer Dong Dong who attracted more than two million livestream viewers, and a digital counterpart of the athlete Eileen Gu hosted via China Mobile's Migu platform. Luo Tianyi, China's first virtual idol created in July 2012, performed at the CCTV Spring Festival Gala in 2021 and was co-opted as a Communist Youth League youth ambassador, amassing more than five million Weibo followers. AYAYI, created by Ranmai Technology and debuted on Xiaohongshu in May 2021, developed international brand partnerships with Guerlain, Louis Vuitton, and Procter & Gamble.

China has pursued an active role in shaping international AI governance norms. Xi Jinping proposed the Global AI Governance Initiative at the third Belt and Road Forum in October 2023, articulating principles including people-centered development, respect for sovereignty, opposition to technological monopolies, and risk-based governance. A China-proposed United Nations General Assembly resolution on AI capacity building was adopted by consensus in July 2024 with support from more than 120 member states including the United States. At the 2025 World AI Conference in Shanghai, China released a 13-point Global AI Governance Action Plan and proposed creating a global AI cooperation organization potentially headquartered in Shanghai. China signed the Bletchley Declaration in November 2023 as one of 28 countries plus the European Union, though it attended but did not sign the Seoul Declaration in May 2024. The first US-China intergovernmental AI dialogue took place on May 14, 2024 in Geneva, and in November 2024, Presidents Biden and Xi agreed that humans rather than AI should make decisions about nuclear weapons. In standards development, China holds a participating membership in ISO/IEC JTC 1/SC 42, whose inaugural meeting was held in Beijing in April 2018, and Chinese experts serve in leadership positions within its working groups. The April 2026 draft regulation on digital virtual humans itself encourages active participation in international rulemaking, exchanges, and cooperation.

The use of digital humans as carriers of Chinese cultural heritage has explicit support in state media and government policy. People's Daily Online reported in September 2023 that digital humans had become new "messengers" of traditional Chinese culture, quoting Liu Guiru of the Chinese Academy of Social Sciences as saying that digital humans can become carriers of Chinese culture. Concrete examples include a digital replica of the Peking Opera master Mei Lanfang, created by the Central Academy of Drama and the Beijing Institute of Technology, which debuted in 2021 and whose New Year's greeting received 50 million views, and Tianyu, a digital avatar inspired by Dunhuang Flying Apsaras and Tang Dynasty figurines. The CPC Central Committee and State Council issued a guideline on national cultural digitalization strategy in May 2022 with an aim to digitalize China's cultural resources through a big data system by 2035, encompassing platforms such as Digital Dunhuang and Panoramic Forbidden City. The 14th Five-Year Plan Cultural Development Plan calls for "digital reproduction" of cultural resources. The government's operative formula for this project is "creative transformation and innovative development" of traditional culture, framing these digital initiatives within the broader concept of excellent traditional Chinese culture rather than as a bridge between ancient philosophy and modern technology.

A growing body of academic literature addresses the intersection of Chinese values and digital human development from multiple angles. Research spans cross-cultural moral judgments about virtual humans, the ethical boundaries of digital human technology in ideological-political education, governance frameworks for ethical and legal risks in virtual human operations, and the demand mechanisms and ethical dilemmas of AI resurrection. Broader studies examine Confucian philosophy's implications for AI governance, the comparison of AI ethics principles between China and the European Union, and the operationalization of AI governance through practices such as data annotation and manual alignment. Research on cultural heritage tourism has begun exploring the relationship between digital human avatars and tourist behavior at heritage destinations. While no single comprehensive work yet unifies the exact intersection of Chinese societal values and digital human development as a standalone field, at least eighteen papers published between 2020 and 2025 collectively address its constituent dimensions, suggesting an emerging interdisciplinary domain with significant room for further scholarly consolidation.

[Apr 2026]