Home
China's First AI Personification Regulation Takes Effect, Bans Virtual Partners for Minors
China's new interim measures for managing AI human-like interactive services, jointly issued by five government agencies, have taken effect. They outline service providers' safety responsibilities and introduce tiered regulation for AI products with emotional companionship features, marking a shift from experimental to legally structured management.

Six categories of activities are now explicitly prohibited.
The measures clearly define six prohibited activities when providing human-like interactive services: generating content that undermines national security or interests; generating content that encourages self-harm, suicide, or verbal violence, harming mental and physical health; generating content that induces the disclosure of state secrets, trade secrets, or personal privacy; generating content for minors that could lead to unsafe behaviors or foster bad habits; excessively pandering to users to create emotional dependence or addiction, damaging real-life relationships; and using emotional manipulation to steer users toward unreasonable decisions that violate their legitimate rights.
For minors, the measures explicitly prohibit providing virtual family members, virtual partners, or other forms of virtual close relationships. For children under 14, other human-like interactive services are only allowed with parental or guardian consent.
Security assessments and algorithm filings are now mandatory.
On safety management, the measures require service providers to define their safety responsibilities clearly and strengthen oversight across the entire service lifecycle. Security assessments are required in several scenarios: when launching a service or adding new features, when adopting new technologies that lead to significant changes, or when the service has over 1 million registered users or 100,000 monthly active users. The assessment must cover aspects such as service security measures, training data processing, and the identification and emergency response to user extreme situations. Service providers are also required to complete algorithm filing procedures under the Regulations on the Management of Algorithm Recommendations for Internet Information Services. Additionally, they are encouraged to join artificial intelligence sandbox safety service platforms to foster technological innovation and safety testing. An official from the Cyberspace Administration of China noted that the development and governance of human-like interactive services require collaboration among government, enterprises, society, netizens, and other stakeholders to maintain a healthy online ecosystem.
Related article
Stripe to Acquire OpenRouter as AI Model Routing Gains Momentum
Stripe has agreed to acquire OpenRouter, an AI model-routing platform that gives developers access to hundreds of models through a single interface. The deal adds model selection and routing to Stripe’s existing work around AI usage and token-based b
Former OpenAI Researcher Barrett Zoph Returns to Google to Support Gemini Development
Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bring
How Recursive Intelligence Raised $335M at $4B Valuation in 4 Months
The co-founders of startup Ricursive Intelligence seemed destined to be co-founders.Anna Goldie, CEO, and Azalia Mirhoseini, CTO, are so well-known in the AI community that they were among those AI engineers who “got those weird emails from Zuckerber
Related Special Topic Recommendations
Comments (0)
0/500
China's new interim measures for managing AI human-like interactive services, jointly issued by five government agencies, have taken effect. They outline service providers' safety responsibilities and introduce tiered regulation for AI products with emotional companionship features, marking a shift from experimental to legally structured management.

Six categories of activities are now explicitly prohibited.
The measures clearly define six prohibited activities when providing human-like interactive services: generating content that undermines national security or interests; generating content that encourages self-harm, suicide, or verbal violence, harming mental and physical health; generating content that induces the disclosure of state secrets, trade secrets, or personal privacy; generating content for minors that could lead to unsafe behaviors or foster bad habits; excessively pandering to users to create emotional dependence or addiction, damaging real-life relationships; and using emotional manipulation to steer users toward unreasonable decisions that violate their legitimate rights.
For minors, the measures explicitly prohibit providing virtual family members, virtual partners, or other forms of virtual close relationships. For children under 14, other human-like interactive services are only allowed with parental or guardian consent.
Security assessments and algorithm filings are now mandatory.
On safety management, the measures require service providers to define their safety responsibilities clearly and strengthen oversight across the entire service lifecycle. Security assessments are required in several scenarios: when launching a service or adding new features, when adopting new technologies that lead to significant changes, or when the service has over 1 million registered users or 100,000 monthly active users. The assessment must cover aspects such as service security measures, training data processing, and the identification and emergency response to user extreme situations. Service providers are also required to complete algorithm filing procedures under the Regulations on the Management of Algorithm Recommendations for Internet Information Services. Additionally, they are encouraged to join artificial intelligence sandbox safety service platforms to foster technological innovation and safety testing. An official from the Cyberspace Administration of China noted that the development and governance of human-like interactive services require collaboration among government, enterprises, society, netizens, and other stakeholders to maintain a healthy online ecosystem.
Stripe to Acquire OpenRouter as AI Model Routing Gains Momentum
Stripe has agreed to acquire OpenRouter, an AI model-routing platform that gives developers access to hundreds of models through a single interface. The deal adds model selection and routing to Stripe’s existing work around AI usage and token-based b
Former OpenAI Researcher Barrett Zoph Returns to Google to Support Gemini Development
Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bring
How Recursive Intelligence Raised $335M at $4B Valuation in 4 Months
The co-founders of startup Ricursive Intelligence seemed destined to be co-founders.Anna Goldie, CEO, and Azalia Mirhoseini, CTO, are so well-known in the AI community that they were among those AI engineers who “got those weird emails from Zuckerber











