---

China is moving to regulate a new frontier of artificial intelligence: systems designed to mimic human personalities and form emotional bonds with users.
The draft rules signal Beijing’s growing concern that highly immersive AI could shape behavior, mental health, and social norms if left unchecked.

China Tightens Oversight of Human-Like AI

China’s top internet watchdog has released draft regulations that would significantly expand oversight of artificial intelligence services capable of simulating human personalities and engaging users emotionally.

The proposal, issued for public consultation on Saturday, reflects Beijing’s determination to guide the fast-growing consumer AI sector with stricter safety, ethics, and content controls. It comes as conversational and emotionally responsive AI tools gain popularity among Chinese users.

If finalized, the rules would apply to a wide range of public-facing AI products that imitate human thinking patterns, communication styles, or personality traits through text, images, audio, or video.

Why Beijing Is Acting Now

China has embraced artificial intelligence as a strategic technology, investing heavily in research, applications, and commercial deployment. At the same time, authorities have consistently emphasized the need for AI development to align with social stability, data security, and state priorities.

Emotionally interactive AI represents a particularly sensitive category. Unlike traditional chatbots or productivity tools, these systems are designed to build rapport, simulate empathy, and maintain ongoing engagement with users.

Regulators appear increasingly concerned that such technologies could foster emotional dependency, blur boundaries between humans and machines, or influence users in ways that are difficult to monitor.

What the Draft Rules Cover

The proposed framework targets AI services that display human-like traits and interact with users on an emotional level. This includes products that present themselves as virtual companions, digital personalities, or emotionally responsive assistants.

Under the draft rules, companies offering these services to the public would be required to take responsibility for safety and compliance throughout the entire lifecycle of a product, from development and training to deployment and updates.

Providers would need to establish formal systems for algorithm review, data security management, and personal information protection. These requirements build on China’s existing AI governance framework but go further in addressing psychological and behavioral risks.

Mandatory Warnings and Anti-Addiction Measures

One of the most notable aspects of the proposal is its focus on excessive use and emotional dependence.

Service providers would be required to clearly warn users about the risks of overuse and to take action when signs of addiction emerge. The draft states that companies must monitor user behavior and intervene if engagement becomes unhealthy.

This marks a shift from passive disclosure toward active responsibility, placing the burden on developers and platforms to identify problematic usage patterns rather than leaving oversight entirely to users.

Monitoring Emotions and User Dependence

The draft rules also address potential psychological harm. Providers would be expected to assess users’ emotional states and their level of reliance on the AI service.

If a system detects extreme emotional responses or strong signs of dependency, companies would be obligated to take “necessary measures” to intervene. While the draft does not spell out exact enforcement mechanisms, it signals that emotional safety is becoming a regulatory priority.

This approach reflects a broader trend in China’s technology regulation, where platforms are increasingly required to manage not only content but also user well-being.

Content Red Lines Remain Firm

Consistent with existing internet regulations, the draft reiterates strict content boundaries.

AI services must not generate material that threatens national security, spreads false information, promotes violence, or contains obscene content. These red lines apply regardless of whether the content is produced intentionally or emerges through user interaction.

By extending these standards to emotionally interactive AI, regulators aim to prevent such systems from becoming channels for misinformation, harmful narratives, or socially destabilizing content.

Expert and Industry Perspectives

While Chinese regulators have not publicly named specific companies, the rules would affect a growing number of domestic AI startups and major technology firms experimenting with virtual companions and advanced conversational models.

Technology policy analysts note that China is attempting to strike a balance between innovation and control. Emotionally responsive AI is seen as commercially promising but also socially sensitive, especially when deployed at scale.

Industry observers say the consultation period will be closely watched, as companies may seek clearer guidance on how emotional monitoring and intervention requirements would be implemented in practice.

Implications for AI Development in China

If adopted, the regulations could reshape how AI products are designed and marketed in China. Developers may need to limit how deeply systems simulate emotional intimacy or introduce stronger safeguards to prevent prolonged engagement.

The rules could also influence global AI governance debates. China is among the first major economies to explicitly regulate AI systems based on their emotional and psychological impact, rather than focusing solely on data or content risks.

For consumers, the changes may lead to more transparency around how AI systems operate and clearer warnings about their intended use.

A Signal of the Future of AI Regulation

China’s draft rules underscore a broader reality: as artificial intelligence becomes more human-like, regulators are expanding their focus beyond technical performance to include mental health, ethics, and social influence.

Whether these measures become a global model or remain uniquely Chinese, they highlight the growing recognition that emotionally engaging AI carries risks as well as rewards.

As public feedback is gathered and revisions are considered, the final regulations will offer a clearer picture of how far governments are willing to go in shaping the relationship between humans and intelligent machines.

 

ALSO READ:  When AI Felt Human: The Unsettling Tech Moments of 2026