China’s New AI Rules Shut Down ByteDance, Alibaba, and Tencent Companion Services

China's-new-AI-rules

China’s new AI rules have pushed ByteDance, Alibaba, and Tencent to remove AI companion features from Doubao, Qwen, and Yuanbao over safety and privacy concerns.

China’s new AI rules shut down ByteDance, Alibaba, and Tencent companion services

China’s new AI rules shut down ByteDance, Alibaba, and Tencent companion services as Beijing increases oversight of chatbots designed to simulate human personalities and emotional relationships. The measures have prompted ByteDance’s Doubao, Alibaba’s Qwen, and Tencent’s Yuanbao to remove or disable features that allowed users to create customised AI personas and interact with them over extended conversations.

The companies are not shutting down their main chatbots. Instead, they are removing specific companion-style tools that regulators believe could encourage unhealthy emotional dependence, expose younger users to harmful content, or create privacy risks.

What China’s rules cover

The new framework, known as the Interim Measures for the Administration of AI Anthropomorphic Interactive Services, was issued by China’s Cyberspace Administration and other government agencies. It came into force on July 15, 2026, after being announced in April.

The rules apply to AI services that imitate human personality traits, thought patterns, or communication styles to support long-term emotional interaction. Regulators are particularly concerned about systems that encourage users to treat an AI as a friend, partner, family member, or replacement for real-world relationships.beckmann+1

The framework also prohibits services from deliberately triggering extreme emotional responses in minors. Providers must introduce safeguards such as usage reminders, anti-addiction controls, rapid exit options, and systems designed to identify signs of unhealthy dependence.

What happened to Doubao, Qwen, and Yuanbao

ByteDance reportedly disabled Doubao’s feature for creating customised AI personas on July 15. Alibaba’s Qwen also removed comparable custom-agent and companion functions, while Tencent’s Yuanbao reportedly withdrew its feature earlier in June.

Reports indicate that some existing user-created agents are being placed in read-only mode before deletion, while certain Qwen configurations may already have been permanently removed. These changes could affect users who spent time building personalised characters, memories, and conversation settings.

The services’ standard functions remain available. Users can still use these platforms for general questions, writing, knowledge search, workplace assistance, and other interactions that do not involve sustained emotional companionship.

Why AI companions are under scrutiny

AI companions have become popular because they offer personal, continuous conversations that can feel more intimate than traditional search or customer-service tools. That same quality has raised concerns about excessive use, emotional manipulation, misinformation, and the collection of sensitive conversation data.

China’s rules also prohibit providers from using private user conversations to train future models without appropriate safeguards. For minors, the restrictions are stricter, particularly around virtual romantic or family-style relationships.

A major test for the AI industry

The shutdowns show how quickly regulation can reshape consumer AI products. Features that once helped companies differentiate their chatbots are now becoming compliance risks, especially when they allow users to create personalities with unrestricted memories and behaviour.

Business assistants, educational tools, knowledge systems, and customer-service bots are generally outside the main restrictions as long as they do not encourage prolonged emotional interaction.

Summary: China’s new AI rules have forced ByteDance, Alibaba, and Tencent to remove companion-style features from Doubao, Qwen, and Yuanbao. The move highlights Beijing’s concerns about emotional dependency, child safety, misinformation, and the privacy implications of long-term AI conversations.

Read Previous

Grok Imagine Image 2.0 Launches With Precision Editing and Clearer AI-Generated Text

Read Next

Redmi Note 17 5G vs Vivo S2: Which New Mid-Ranger Should You Buy?