China unveils draft rules to regulate humanoid artificial intelligence

Regulation Focused on Transparency, User Protection, and Human Simulation

The Cyberspace Administration of China (CAC) has released draft regulations for artificial intelligence (AI) services capable of simulating human interaction, a move aimed at safeguarding national security and society and ensuring the ethical use of intelligent technologies. This development, announced on December 27, 2025, marks a turning point in how governments approach advanced AI and its interactive applications.

Scope of Censorship and Content Red Lines

The draft targets systems capable of simulating human personality traits, thought patterns, and emotional interactions through text, images, audio, or video. The regulations will be implemented within the territory of the People’s Republic of China, under the central supervision of the China Internet Information Office, with gradual local censorship based on the size and impact of each service.

The draft outlines a set of strict content red lines, including:

• Violent, pornographic, gambling, or incitement to crime content.
• Threatening national security or national unity.
• Glorifying suicide or self-harm.
• Defaming or abusing individuals. Emotional manipulation or deceiving the user into making irrational decisions.
• Protecting users and vulnerable groups
• The new rules focus specifically on reducing emotional dependency and digital addiction, requiring service providers to:
• Intervene immediately when a user exhibits dangerous tendencies.
• Send an automatic warning after two hours of continuous use.
• Provide clear and continuous reminders that the interaction is with an AI system.
• The regulations also include special provisions to protect vulnerable groups:
• Minors: Require explicit parental consent and provide controls covering time, content, and spending.
• Elderly individuals: Mandate the establishment of emergency contacts and prohibit the simulation of relatives or personal relationships to prevent emotional manipulation.

Data, assessments, and penalties

The draft imposes mandatory security and ethics assessments on large-scale services and prohibits the use of sensitive data to train models without explicit consent. It also guarantees users the right to delete interaction data, with penalties ranging up to suspension or complete service termination for violations.

One of the comments accompanying the draft stated:

“Requiring AI service providers to be transparent and conduct security and ethical assessments is a crucial step towards building trust in these technologies. The future lies not only in the power of AI, but also in how responsibly we use it.”

Practical Impact and STEMpire Response

For entrepreneurs, training managers, and decision-makers in the region, these guidelines underscore that AI governance and ethics are no longer theoretical issues, but rather essential operational requirements. Organizations that adopt interactive intelligent systems are now required to place transparency, user protection, and risk management at the heart of their digital strategies.

At STEMpire, we see these global developments as an opportunity to enhance organizational readiness through practical training and tailored programs in AI governance, compliance, and digital ethics.

STEMpire’s Recommendations for Organizations:

1. AI Governance: Develop clear internal policies that ensure transparency and accountability and define the roles of central and local oversight.
2. Compliance and Protection of Vulnerable Groups:Training teams on compliance requirements and designing user experiences that are sensitive to the needs of minors and the elderly.
3. Security and Ethical Assessment: Integrating periodic reviews of intelligent systems to ensure the reduction of bias and behavioral risks.
4. Sensitive Data Management: Establishing strict data use protocols and guaranteeing the user's right to delete interaction data.
   
   To inquire about how to practically implement these ethical and organizational standards within your organization, contact STEMpire for a consultation or a customized training program.