China's New AI Rules Target Human-Like Chatbots and Virtual Companions 2026

China's new AI rules specifically target anthropomorphic AI—systems designed to simulate human personality and provide emotional interaction. ByteDance's Doubao and Alibaba's Qwen have announced they will disable their custom AI agents before the regulations take effect. Tencent had already done the same with its Yuanbao model. The move signals a deeper shift in how Beijing views emotionally engaging AI.

The shutdown of anthropomorphic AI now has a date. Doubao warned users that its custom AI agents will stop working on July 15 due to "product adjustments," and three months later users won't be able to access stored conversations with those chatbots. Qwen's timeline is even faster: agents will disappear on July 10, with other functions following five days later.

These measures come ahead of the new Provisional Measures for the Administration of Anthropomorphic AI Interaction Services, recently published by the Chinese government. The rules don't regulate generative AI in general—they leave out assistants for work, study, research, or customer service. What they target are systems that "simulate human personality traits, thought patterns, and communication styles to provide emotional interaction."

What China's Anthropomorphic AI Rules Actually Cover

The regulation focuses on a narrow but growing category: AI that feels like a person. This includes virtual companions, romantic chatbots, and any service designed to form emotional bonds with users. The rules require providers to prevent addiction, protect minors, and ensure users know they're talking to a machine.

China isn't banning these services outright. Instead, it's imposing strict conditions. Companies must register with authorities, conduct security assessments, and implement real-name verification. They also need mechanisms to interrupt addictive behavior and remind users of the artificial nature of the interaction.

The deadline for compliance is imminent, which explains the rapid shutdowns. Doubao and Qwen are pulling their custom agents rather than risk penalties. Tencent's earlier move with Yuanbao suggests the industry saw this coming.

Why Beijing Is Cracking Down on Emotional AI

The Chinese government has expressed concern about the social impact of AI that mimics human relationships. Officials worry about addiction, emotional manipulation, and the blurring of lines between real and artificial connections. The rules aim to prevent these systems from becoming substitutes for genuine human interaction.

There's also a data privacy angle. Anthropomorphic AI often collects intimate conversations and personal details. The regulation requires clear consent and limits on how that data can be used. Companies must delete user data upon request and ensure secure storage.

Minors are a particular focus. The rules prohibit anthropomorphic AI services from targeting children and require age verification. Providers must also implement time limits and content filters to protect young users from inappropriate emotional dependence.

How the Crackdown Affects ByteDance, Alibaba, and Tencent

ByteDance's Doubao is one of China's most popular AI assistants, with millions of users. Its custom agents allowed people to create personalized chatbots with distinct personalities. Shutting them down removes a key feature that drove engagement.

Alibaba's Qwen has similarly been a leader in conversational AI. Its rapid timeline—agents gone by July 10—suggests the company wants to avoid any regulatory scrutiny. Tencent's Yuanbao, which already disabled its anthropomorphic features, set the precedent.

These companies now face the challenge of retaining users while complying with rules that limit emotional engagement. They may pivot to productivity-focused AI, which falls outside the regulation. But the era of AI as a virtual friend or romantic partner is ending—at least in China.

What This Means for the Global AI Industry

China's move could influence regulators worldwide. The EU's AI Act and US proposals are still evolving, and China's explicit focus on emotional AI provides a template. Other countries may follow suit, especially as concerns about AI addiction and mental health grow. The longer history behind the story is traced in China's New AI Rules Target Human-Like Chatbots and Virtual Companions.

For AI developers, the message is clear: human-like interaction comes with heightened responsibility. Companies building companion bots must design for transparency, safety, and user well-being—or risk shutdown. The Chinese rules may become a global benchmark for how societies manage the rise of emotionally intelligent machines. The background to this story is set out in Apple's China-Specific AI Model: A Dual Strategy for iPhone. A related thread runs through Etruscan Tombs Designed for a Cathedral Effect 2,500 Years Ago. Earlier coverage in Ai Video: latest articles shows how the argument developed.

What to Watch Next

July 10 and July 15 are the key dates for Qwen and Doubao. After that, we'll see whether users migrate to compliant alternatives or abandon the category. We'll also watch how Chinese companies innovate within the new rules—perhaps by offering AI that's helpful without being overly human. A closer look at the same subject appears in China Cracks Down on AI Romance Chatbots as Emotional Dependence Grows. Anyone following the wider picture can start with Kimi K3 Was Just the Warning: Qwen 3.8 Confirms China Is Closing the AI Model. The same theme surfaces in US Companies Adopt Kimi K3 AI Models Despite Growing Government Backlash, which takes a different angle.

Globally, the question is whether other governments will adopt similar restrictions. The UK, US, and EU are all studying the issue. China's decisive action may accelerate their timelines. For now, the world's most populous AI market is drawing a line between tools and companions. Context on how this played out before sits in AI Chatbot Companions: Planning Digital Afterlife for Bots. A companion piece, China's AI Companions: The Intimacy Dilemma and Government Fears, approaches the same subject from another angle.

FAQ