The student sitting in a lecture hall at the Hong Kong University of Science and Technology’s Guangzhou campus might be listening to Albert Einstein explain the theory of relativity, or hearing John Nash walk through the mathematics of game theory. Neither man is alive. Both are present, as AI avatars, built from video-generation models and large language models, capable of delivering lectures and answering student questions in real time.
This is not a novelty experiment. It is part of a systematic deployment of digital humans across Chinese universities, hospitals, and social platforms that has grown faster than anyone anticipated and faster, certainly, than the regulators tasked with governing it.
A 1.3 Million-Company Industry
China now has more than 1.3 million companies offering digital avatar services, according to a 2024 government report cited in a Nature feature published this week. The scale of that figure is striking: it means that digital human services have become a mainstream industry rather than a niche technology, with applications spanning healthcare, education, retail, and social companionship. The Cyberspace Administration of China’s new rules governing virtual digital humans took effect on July 15, marking the first comprehensive national regulatory framework for a sector that has been operating largely without formal oversight.
AI Doctors and Virtual Hospitals
The healthcare deployment is among the most consequential. Ant Group, the Hangzhou-based fintech giant, launched an AI health application last June that features digital versions of more than 1,000 real-world physicians. These avatars can answer patient questions, recommend diagnostic tests, and propose treatment plans, though human physicians retain final decision-making authority over all care. By January 2026, the app had reached 30 million active monthly users, a figure that places it among the most widely used AI health tools worldwide.
Separately, a team at Tsinghua University has been running trials of Agent Hospital, a virtual hospital simulation comprising AI doctors, nurses, and patients, at eight hospitals across China since November 2025. The system provides human physicians with personal digital avatars to assist with administrative tasks, note-writing, and clinical recommendations.
Einstein in the Classroom
The education sector has moved at a similar pace. Computer scientist Pan Hui and his team at HKUST-Guangzhou began using AI avatars for teaching in 2024, creating ten digital lecturers that delivered parts of a social media course. The roster included avatars modeled on Einstein and Nash alongside fictional characters of different ages and nationalities.
Hui says the initiative was driven by student demand for more guest lecturers, a resource constraint that AI avatars can eliminate entirely. The team currently uses a mix of interactive and pre-recorded avatar formats, both computationally intensive to generate. As the underlying technology improves, Hui envisions AI avatars evolving into personal tutors available around the clock.
AI Proxies, Dating, and the Social Layer
The social dimension of China’s avatar economy is perhaps the most culturally distinctive. A networking service called Elys allows users to create AI clones of themselves that interact with other avatars, commenting, liking posts, and building social connections, without the human user needing to be online. Some users have repurposed the service as a dating platform, using their AI proxies to identify compatible matches before engaging directly. Mengjiao Yin, a researcher studying AI and digital humans at Wuxi Taihu University, notes that while no scientist avatars are yet working independently in laboratories, she considers that milestone “not far off.”
Regulation Racing to Catch Up
The regulatory framework that took effect this month reflects both the government’s enthusiasm for the technology and its awareness of the risks it creates. The CAC rules encourage the deployment of AI avatars across Chinese society and aim to accelerate collaboration between academia and digital human developers, a notably permissive posture compared to the more cautious approach taken by European regulators.
At the same time, the rules address concerns about deepfakes, unauthorized identity cloning, and the potential for AI avatars to deceive or manipulate users. EastFrontier has previously covered China’s evolving digital human regulatory landscape, including the April 2026 draft rules that preceded this month’s final framework.
Yang Liu, a computer scientist at Tsinghua University, has described AI avatars as likely to “profoundly transform” daily life and reshape every aspect of human society. “Acting as our proxies and assistants, they will take on a wide range of tasks and greatly boost productivity,” he told Nature. He also cautioned that the technology still needs to improve, and that equal effort must go into developing ethical standards for how AI avatars operate, a balance that China’s new regulations attempt to strike, though critics argue the framework prioritizes deployment over protection.
The broader context is China’s national AI strategy, which prioritizes the widespread societal adoption of AI and embeds it in the current five-year plan for social and economic development. Digital humans are not a side project in that strategy, they are a central mechanism for scaling AI’s impact beyond the technology sector and into the daily lives of ordinary citizens.
Whether that scaling proceeds responsibly will depend on how rigorously the new CAC rules are enforced, and whether the regulatory framework can keep pace with a sector that has already outgrown every previous attempt to govern it. As EastFrontier reported on China’s AI governance push at WAIC 2026, Beijing is simultaneously accelerating AI deployment and tightening the rules that govern it, a tension that will define the next phase of China’s AI story.