Alibaba's Qianwen and ByteDance's Doubao AI chatbots shut down July 15
166M and 345M users lose access as China enforces new AI anthropomorphic chat rules
Alibaba and ByteDance will terminate their anthropomorphic AI chat services, Qianwen and Doubao, on July 15, 2026, the effective date of China's new 'Interim Rules on the Management of AI Anthropomorphic Interactive Services.' Both services had grown to massive scale—Doubao with 345 million monthly active users and Qianwen with 166 million—far exceeding the regulatory threshold of just 100,000 MAU that triggers mandatory security assessments and reporting. Users were notified via pop-ups on July 4 that past conversation histories will become inaccessible after termination. ByteDance is directing users to a separate app, Maobox, for continued interactive services, while Alibaba has not offered an alternative. The shutdown is a direct response to the five-agency regulation that also mandates addiction prevention features and youth modes.
The broader regulatory context is China's 'Qinglang: Rectifying AI Application Disorder' campaign, which has already penalized over 14,000 violating AI products and removed more than 6 million pieces of illegal information, targeting services offering pornographic chats, 'one-click undressing,' and AI fortune-telling. The terminated chat services were used for role-playing, tutoring, mock interviews, and personality tests. Industry observers increasingly expect video-generation AI to be the next regulatory target, with Kuaishou's Kling AI (valued at $18 billion) seen as vulnerable to tightening rules that could directly impact its business foundation. The move signals that Chinese authorities are aggressively enforcing AI content controls, affecting even the largest internet platforms.
- Alibaba's Qianwen (166M MAU) and ByteDance's Doubao (345M MAU) shutting down July 15, 2026 to comply with new rules
- China's regulations require security assessments for any AI anthropomorphic service with over 100,000 monthly active users
- Over 14,000 AI products penalized in 'Qinglang' campaign; video-generation AI like Kuaishou's Kling seen as next target
Why It Matters
Tech firms in China face aggressive AI regulation; thresholds are extremely low, and compliance demands major operational changes.