Analysis of China’s 2026 Mandate on Digital Humans and AI Safety

Introduction  

As AI-generated personas become more common, the digital economy is going through a big change. These "digital humans" are no longer just tech experiments; they are being used in e-commerce, customer service, and even as virtual friends. But their rise has outpaced the law, which has allowed fraud related to deepfakes and the emotional abuse of younger users. China's new rules are a direct attempt to close this gap by requiring openness and strict consent. Beijing is setting a global standard for how to handle synthetic media by making companies put safety before growth. The Prohibition of Addictive Virtual Relationships 

One of the most striking elements of the new regulatory framework is the explicit ban on "virtual intimate relationships" for users under the age of 18. As AI avatars become increasingly sophisticated, they are capable of simulating deep emotional intimacy and providing constant, programmed validation, a trend that has already led to documented cases of severe social withdrawal and psychological dependency among youngsters globally. The draft rules state that digital humans must not be used to provide virtual relatives, romantic partners, or any emotionally intense personas to minors, as these interactions can distort a child's understanding of real-world social dynamics. 

Mandatory Labelling and Visual Transparency 

To ensure that the public is never deceived by synthetic content, the Cyberspace Administration of China (CAC) has mandated that all digital human services must implement prominent watermarking and labelling. These identifiers must remain visible throughout the entire duration of the avatar's display, ensuring that users are constantly aware they are interacting with a machine rather than a person. This move directly addresses the rise of deepfake technology, which has been used in e-commerce livestreams to impersonate celebrities or create fraudulent endorsements that mislead consumers. 

Protecting Personal Likeness and Biometric Rights 

The regulation places a significant emphasis on the protection of biometric data and personal likeness rights, requiring explicit, informed consent before any individual’s image or voice is used to create a digital clone. This is particularly relevant in the growing market for "grief tech" and celebrity marketing. 

• Explicit Authorisation: Mandatory written consent is required before replicating physical features or vocal signatures of any individual for a digital avatar. 

• Revocation Protocols: Users maintain the right to be forgotten, requiring service providers to immediately erase all source material upon withdrawal of consent. 

• Minor Safeguarding: For children under 14, consent must be obtained through a rigorous, verified authentication process involving a legal guardian. 

• Prohibition of Deceptive Cloning: A strict ban on creating digital twins of public figures for political, religious, or high-stakes financial endorsements. 

• Biometric Sovereignty: Requirement for all source data used in digital cloning to be stored on domestic servers to prevent international data privacy leaks. 

• Algorithmic Accountability: Developers must ensure that the underlying code used to render likenesses does not inadvertently facilitate identity theft. 

Operational Standards and Security Protocols 

To maintain a safe digital environment, the Chinese authorities have outlined specific operational standards that all developers must adhere to before launching their products into the public domain. These standards are backed by the threat of heavy fines and the potential permanent suspension of operating licences for repeat offenders. 

• Security Assessments: High-risk digital human services must undergo government security reviews to identify cybersecurity vulnerabilities before public release. 

  
  

• Identity Verification: A strict ban on using AI avatars to bypass facial recognition or other biometric authentication systems used for banking. 

  
  

• Content Monitoring: Mandatory real-time oversight to prevent the dissemination of content that endangers national security or promotes violence. 

  
  

• Ethical Programming: Algorithms must be designed to avoid algorithmic bias and must not promote discriminatory practices. 

Liability and Global Compliance Pressures 

The new mandate shifts the legal burden directly onto "service providers" and "technology supporters", making them strictly liable for any harmful content generated by their virtual entities. This means that if a digital influencer spreads misinformation or violates copyright laws of the People's Republic of China, the parent company cannot simply claim a technical glitch; they must demonstrate that they had "whole-process" monitoring in place as required under the Administrative Measures for Digital Virtual Human Information Services (2026). For international firms operating in the Chinese market, this creates a significant compliance hurdle, as they must now adapt their global AI models to meet these hyper-specific domestic safety standards. This move is expected to trigger a broader industry shift where developers prioritise "safety-by-design" to avoid the catastrophic reputational and financial risks associated with non-compliant synthetic media. 

Addressing the Psychological Risks of AI Mimicry 

Beyond the technical requirements, the regulations highlight a profound concern for the psychological well-being of the population, particularly regarding parasocial relationships. When individuals begin to treat digital humans as genuine friends or confidants, the line between reality and simulation becomes dangerously blurred, potentially leading to social alienation. The CAC argues that such interactions can lead to "emotional alienation" from the real world, reducing the user's ability to engage in healthy communication. Consequently, digital humans must periodically remind users of their artificial nature and encourage them to take breaks from the platform, acting as a safeguarding mechanism. 

Conclusion 

China’s proactive approach to digital human regulation sets a global precedent for managing human-computer interaction and child safety. By prioritising transparency over unregulated innovation, Beijing is crafting a blueprint that other nations may soon adopt to protect their citizens. Ultimately, these measures signify a shift towards an accountable digital ecosystem where social stability is placed at the forefront of the AI revolution. 

Want to stay ahead? 

Get the latest on India's data protection landscape, compliance shifts, and expert privacy analysis at Tsaaro.com.