China is moving to strengthen regulatory oversight of its rapidly growing artificial intelligence (AI) "digital human" industry, which produces highly realistic avatars used in social media, marketing, and increasingly in private, personal contexts. The Cyberspace Administration of China (CAC) recently released draft rules aimed at governing the development and deployment of these AI-generated digital humans, seeking to curb potential risks such as harm to minors, threats to social stability, and the unauthorized creation of avatars resembling real individuals.
The draft regulations, published this month and open for public comment until early May, require clear labeling of AI-generated digital human content. They explicitly prohibit the use of personal data to create deepfake clones without consent, aiming to safeguard individual privacy and prevent deceptive practices. Under the proposed framework, violations could attract fines ranging from 10,000 yuan ($1,460) to 200,000 yuan.
China’s digital human sector has seen rapid growth, with state media reporting the industry’s valuation at approximately 4.1 billion yuan (around $600 million) in 2024, an increase of 85% from the previous year. These avatars, which combine lifelike appearance and fluid motion, have become a common presence on Chinese social media platforms, often serving as brand promoters or interactive interfaces.
Among the emerging uses of digital humans is the creation of AI avatars that simulate deceased individuals. Zhang Xinyu, a resident of Liaoning province, formed an avatar of her late father through a company called Super Brain, enabling her to engage in online conversations with a digital recreation of him. Zhang described the experience as emotionally uplifting, though she acknowledged some critics view such AI interactions as potentially hindering the grieving process.
Super Brain’s founder Zhang Zewei recognizes the necessity of regulatory measures, viewing them as a means to balance industry innovation with ethical standards. The company has attracted attention for its hyper-realistic avatars, including one that deeply moved an elderly woman recently when she unknowingly spoke to a lifelike avatar of her deceased son on what appeared to be a video call. This case sparked significant social media debate over the morality and implications of AI-generated human likenesses.
Regulators emphasize that digital humans must not produce content that endangers national security, incites subversion, or threatens social harmony. Special provisions also seek to protect minors by banning services that would create virtual intimate relationships with children, reflecting broader concerns about the societal impacts of AI.
This latest move continues China’s pattern of cautious technology governance, adopting a "develop first, then regulate" approach. Analysts suggest that enforcement reflects not only technological considerations but also efforts to maintain political stability and digital sovereignty. Earlier CAC actions targeted abuses such as deepfake impersonations of public figures in e-commerce livestreams, which authorities said undermined trust in online ecosystems.
As China navigates the challenges posed by AI-generated digital humans, the draft rules represent an attempt to harness innovation while addressing legal, ethical, and social risks associated with this evolving technology.