On July 15, the Interim Measures for the Management of Human-like Interactive Services of Artificial Intelligence will come into effect. Two mainstream AI applications in China have already responded. ByteDance's "Doubao" and Alibaba Group's "Tongyi Qianwen" are preparing to remove customizable human-like agent features, becoming the first major platforms to adjust their product forms before the new regulations take effect.

Robot AI Artificial Intelligence 222

Doubao sent a notice to users in the evening on Friday, stating that its agent function will be discontinued starting July 15, citing "product function adjustments." The notice stated that data related to this feature will be handled according to the company's privacy policy starting October 15, and users will no longer be able to view or recover related content within the app. Tongyi Qianwen is more urgent; its "human-like interactive agents and user-created agent functions" will be disabled on July 10, and the broader "Tongyi agent functions and services" will be fully discontinued on July 15, after which users will no longer be able to access existing agent settings or historical conversation records.

New regulations target "continuous emotional interaction," with three major giants retreating

Previously, both applications provided agent pools created jointly by the platform and users, allowing users to set fixed personas and tones for AI assistants, creating named assistants, learning tutors, role-playing characters, or emotional companions. The timing of the functional adjustments coincides closely with the implementation of the regulatory measures. The Interim Measures for the Management of Human-like Interactive Services of Artificial Intelligence released in April clearly include services that "simulate human personality traits, thinking patterns, and communication methods, and provide continuous emotional interaction" within the scope of regulation, but exclude customer service robots, knowledge question-answering tools, workplace intelligent assistants, and educational and research tools. Regulatory documents mention risks such as the spread of extreme ideas, privacy leaks, harm to users' physical and mental health, and resulting dependency or addictive behaviors.

It is worth noting that Tencent has already removed similar agent functions from its AI assistant application "Yuanbao" in June. So far, the three major technology giants in China have shown a collective trend of retreating in human-like agent functions. From the regulatory signal, China intends to incorporate AI agents into future productive infrastructure while tightening control over human-like companion agents that may establish emotional or quasi-social relationships. Between May and June this year, regulators also successively issued guidance on orderly development of AI agents and national standards for interoperability, establishing unified systems in terms of identification, authorization, connectivity, and traceability.

User backlash: Emotional support disappears overnight

The implementation of the policy has triggered some user backlash. A Weibo user @ Doubao official account said, "Why are we discontinuing the agent? They have always been our emotional support." The user also complained that long-term accumulated chat records and emotional connections are difficult to fully export or transfer through existing methods, and the lack of a smooth data migration path makes this adjustment feel more abrupt. When AI companions become a spiritual reliance for some users, the balance between regulation and products obviously needs more buffer and transition time.