China's Anthropomorphic AI Rules Force Major Tech Giants to Disable Custom Agents

In a significant move reflecting the accelerating pace of global AI regulation, Chinese tech giants ByteDance and Alibaba are in the process of disabling customized humanlike AI agent features within their popular applications, Doubao and Qwen, respectively. This action comes in direct response to China's new "Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interaction Services," which are set to take effect on July 15, 2026. ByteDance's Doubao, a ChatGPT-like conversational bot, informed users that its agent feature would go offline on July 15, with associated data handling adjustments by October 15. Similarly, Alibaba's Qwen announced that its humanlike interactive agents would be disabled on July 10, with broader agent functions ceasing by July 15. These rules, issued in April, target AI services that simulate human personality traits and interaction styles. This development is a stark reminder for AI practitioners, particularly those involved in product development and market strategy, that regulatory compliance is no longer a peripheral concern but a core operational challenge. The immediate impact on established product features and user experiences highlights the imperative for agility and foresight in AI development. For companies operating in or targeting markets with evolving AI governance, this serves as a critical case study: regulatory shifts can necessitate rapid and fundamental changes to core functionalities, potentially disrupting user engagement, development roadmaps, and competitive positioning. Practitioners must proactively monitor global AI policy and build systems designed for regulatory adaptability. This regulatory intervention by Beijing is part of a broader, concerted effort to establish a comprehensive and granular regulatory framework for its rapidly expanding AI sector. China has consistently demonstrated a proactive stance in AI governance, often leading the way in areas such as content moderation, data security, and the ethical implications of AI's ability to mimic human behavior. This trend is not isolated; it mirrors a global movement where governments are transitioning from broad ethical guidelines to specific, enforceable laws. The European Union's AI Act, for instance, also includes stringent provisions for high-risk AI systems and mandates transparency, reflecting a shared international concern about the societal impact of advanced AI capabilities. The focus on anthropomorphic interaction specifically underscores concerns about user manipulation, misinformation, and the blurring lines between human and AI interaction. In practice, this means that AI developers and product teams must now integrate legal and policy expertise directly into their development lifecycle. The rapid timeline for these changes—rules issued in April, with enforcement commencing in July—underscores the need for extreme agility. Companies should anticipate similar regulatory pressures in other jurisdictions, particularly as AI models become more sophisticated in simulating human characteristics. This could involve redesigning user interfaces to clearly distinguish AI from human interaction, implementing stricter content and behavior filters for AI agents, or even strategically removing features deemed non-compliant. The trade-off between innovative, humanlike AI experiences and regulatory adherence will become an increasingly central challenge, demanding robust governance frameworks and continuous adaptation from all stakeholders in the AI ecosystem.