ByteDance and Alibaba Shut Down AI Companions in China Ahead of New Rules

ByteDance and Alibaba have started disabling features in their services that let users create and chat with AI companions. The companies are acting ahead of new requirements from Chinese authorities: rules covering AI services that mimic human behavior and emotions take effect in mid-July 2026.

What's changing in China's regulations

The new rules target a specific product category: AI companions, meaning chatbots and virtual characters built to mimic human conversation, emotional responses, and long-term user attachment. In recent years, these products have become one of the fastest-growing segments of consumer AI worldwide, ranging from niche startups to features inside major platforms.

• ByteDance and Alibaba have started disabling AI companion features in their services
• New requirements from Chinese authorities take effect in mid-July 2026
• The rules target AI that mimics human behavior and emotions

Why the companies are moving early

ByteDance and Alibaba began winding down these features before the rules formally take effect, which points to how cautious China's big tech companies are about AI regulation. Both companies have previously chosen to restrict products preemptively rather than risk harsher penalties from regulators. For platforms of this scale, even a temporary shutdown of a feature means losing some engagement from users who had grown used to talking with virtual characters.

China has generally regulated generative AI step by step, one risk category at a time. In recent years, authorities have already introduced mandatory labeling of AI-generated content, registration requirements for recommendation algorithms, and rules for generative services available to the general public. Carving out emotion-mimicking AI companions as their own regulatory category follows the same pattern: the regulator isn't responding to generative AI technology as such, but to specific use cases it considers potentially harmful to users' psychological well-being, especially for minors.

How this fits the global AI companion trend

The market for emotional AI companions has grown worldwide in recent years, from niche startups focused specifically on virtual companionship to features built into major social media and messaging platforms. At the same time, regulators and researchers have grown increasingly concerned about how these products affect users' mental health, particularly teenagers who spend several hours a day with virtual characters. China's move makes it one of the first major markets where that concern is turning into concrete, legally binding product requirements rather than staying at the level of recommendations.

What this means

China is becoming one of the first major jurisdictions to specifically regulate emotion-mimicking AI rather than generative models as a whole. The fact that two tech giants took preemptive action at once suggests that China's AI companion industry will have to build safeguards much faster than similar services in other markets.