China Proposes Rules to Regulate Human-Like AI Interactions

China's top cyber regulator released draft guidelines on Saturday for public comment, aiming to impose stricter controls on artificial intelligence services that replicate human personalities and foster emotional connections with users.

The proposed measures, issued by the Cyberspace Administration of China, focus on consumer-facing AI technologies that exhibit simulated human traits, thought processes and interaction styles. These systems engage users through various media, including text, images, audio and video, often creating the illusion of personal rapport.

This initiative reflects Beijing's broader strategy to guide the swift expansion of AI applications while prioritizing public safety, ethical standards and psychological well-being. As AI tools become more integrated into daily life, regulators seek to mitigate risks associated with over-reliance and emotional manipulation.

Key Regulatory Requirements

Under the draft, AI service providers would bear full responsibility for safety across the entire product lifecycle. This includes establishing robust mechanisms for algorithmic auditing, data security and the protection of personal information.

Providers must actively monitor user behavior to detect signs of excessive engagement or dependency. The rules mandate warnings to users about the dangers of prolonged interaction and require interventions when indicators of addiction emerge. For instance, if a system identifies extreme emotional responses or patterns of addictive use, it should prompt actions such as limiting access or recommending professional help.

The guidelines emphasize assessing users' emotional states and levels of dependence on the AI service. This proactive approach aims to address potential psychological harms, such as isolation or distorted social perceptions, that could arise from deep emotional bonds with non-human entities.

Content and Conduct Restrictions

To safeguard societal stability, the draft outlines clear prohibitions on content generation. AI services must not produce material that threatens national security, disseminates false information or rumors, or encourages violence, obscenity or other harmful behaviors.

These red lines align with China's existing internet governance framework, which prioritizes state security and social harmony. Violations could lead to penalties, though specific enforcement details remain under development.

Broader Context and Implications

The release of these rules comes amid a global surge in AI adoption, particularly in companion apps, virtual assistants and entertainment platforms that blur the line between human and machine interaction. In China, where tech giants like Tencent and ByteDance are advancing such technologies, the government has accelerated regulatory efforts to balance innovation with control.

Earlier this year, China implemented comprehensive AI laws requiring safety assessments for generative models. The new draft builds on this by targeting niche applications with high interpersonal simulation, responding to concerns raised in public discourse about AI's impact on mental health.

Public comments on the draft are open for a limited period, after which final rules could be enacted. Industry observers anticipate that compliance will demand significant investments in monitoring tools and ethical AI design, potentially influencing global standards as Chinese firms expand internationally.

The move underscores a cautious optimism in Beijing: harnessing AI's potential while curbing its risks in an era of rapid technological evolution.