AI-generated “digital humans” are moving into classrooms and other public-facing roles in China as regulators work to establish rules governing how these increasingly human-like systems can be used.
By The Grey Ghost
WHAT’S HAPPENING
AI-generated avatars in China are being developed to look, speak and interact like real people.
At the Hong Kong University of Science and Technology in Guangzhou, students can attend lectures delivered by AI avatars modeled on figures including Albert Einstein and mathematician John Nash. The avatars can also respond to questions from students. Nature reports that digital people are also being developed for roles involving healthcare and companionship. (Nature)
China’s Cyberspace Administration has separately proposed regulations specifically covering “digital humans,” defined as virtual figures that simulate human appearance and can possess voices, behaviors, interactive abilities or personality characteristics. (China Anti-Corruption Agency)
WHY IT MATTERS
China is beginning to establish legal boundaries around AI systems designed not simply to answer questions, but to appear and interact as human beings.
Under the proposed digital-human rules, AI avatars would have to be visibly identified as digital humans. The proposal also addresses how personal information can be used to create avatars, how digital humans are used for identity verification, and how providers monitor safety risks. (China Anti-Corruption Agency)
Separate Chinese regulations covering AI services that simulate human personalities and sustained human-like interaction took effect on July 15, 2026. (China Anti-Corruption Agency)
WHO BENEFITS
Universities and educators can use interactive digital instructors to supplement traditional teaching.
Businesses and public-service organizations gain another way to provide automated, human-like interaction at scale.
AI developers gain a growing market for systems combining generative AI, speech, animation and interactive avatars.
Users gain access to AI interfaces that communicate through a recognizable human form rather than a traditional text box.
WHO LOSES
Companies relying on AI avatars that are difficult to distinguish from real people could face additional disclosure requirements.
Under China’s proposed rules, digital-human services would be required to continuously display an identification indicating that the figure is digital. The proposal also restricts unauthorized use of personal information and the use of digital humans to circumvent identity-verification systems. (China Anti-Corruption Agency)
Providers operating in regulated areas such as healthcare and finance would also remain subject to existing rules governing those industries. (China Anti-Corruption Agency)
WHAT HAPPENS NEXT
China now has two related regulatory tracks developing around human-like AI.
Rules governing AI anthropomorphic interactive services became effective July 15, 2026. Meanwhile, the broader regulations specifically governing digital human information services were released as a draft in April and, as of August 7, have not been published by the Cyberspace Administration as final regulations. (China Anti-Corruption Agency)
The distinction matters as AI avatars expand beyond entertainment and begin appearing in education, healthcare, customer interaction and other roles traditionally performed by people.
