ByteDance’s Doubao and Alibaba’s Qwen apps will disable their humanlike and user-created AI agents by mid-July, complying with China’s upcoming regulations targeting AI services that simulate sustained emotional interactions, marking a significant shift in the country’s fast-evolving AI landscape.
- Humanlike AI agents disabled by July amid new regulatory rules
- China targets AI with sustained emotional interaction in new policy
- Users lose access to custom agent features and conversation histories
What happened
ByteDance’s Doubao and Alibaba’s Qwen, two leading consumer AI applications in China, announced they will disable their customized humanlike AI agent features in July. Doubao’s agent services will go offline on July 15, while Qwen is disabling humanlike interactive agents and user-created agents starting July 10, with full agent function suspension by July 15. After the shutdowns, users will no longer be able to access prior conversations or agent settings.
Both apps had allowed users to create personalized AI assistants with specific personas, voices, and functions tailored to tasks such as tutoring, companionship, and role playing. The takedown coincides with China’s new Interim Measures for Artificial Intelligence Anthropomorphic Interaction Services, set to take effect on July 15. These measures regulate AI systems that simulate human traits in ongoing emotional interactions.
Why it matters
The new regulations reflect Beijing’s cautious stance on AI technologies capable of sustained emotional engagement, citing concerns about extremist content, privacy breaches, mental health impacts, and addiction risks. While other types of AI tools, like customer service bots and educational assistants, remain exempt if they do not involve ongoing emotional interaction, humanlike AI agents face increased scrutiny.
China’s regulatory framework signals a broader effort to balance innovation with safety by emphasizing AI agent safety, practical uses, and standardization. National standards recently released also aim to make AI agents identifiable, authorized, and traceable, which could transform how these systems operate within China’s tech ecosystem and overall market.
What to watch next
Industry observers will be monitoring how Chinese tech giants adapt to the new rules and whether similar restrictions extend to other AI applications beyond chatbots. The regulatory framework’s focus on control and traceability could inspire innovations around transparency and compliance but may also limit user engagement with highly personalized AI companions.
User backlash highlights the emotional role AI agents have begun to play, indicating a demand for seamless data portability or alternatives that comply with regulations while preserving user experience. How Beijing and companies address these tensions will shape the trajectory of consumer-facing AI development in China and potentially offer a model for other jurisdictions considering emotional AI regulation.