Skip to content
gentic.news — AI News Intelligence Platform
Connecting to the Living Graph…

Listen to today's AI briefing

Daily podcast — 5 min, AI-narrated summary of top stories

ByteDance and Alibaba AI agents disabled by China's July 15 ban, affecting Doubao and Qwen services

China's July 15 AI Agent Ban Hits ByteDance, Alibaba

China's July 15 anthropomorphic AI rules force ByteDance and Alibaba to disable custom agents, exempting utility bots while targeting emotional companions.

·1d ago·3 min read··31 views·AI-Generated·Report error
Share:
Source: scmp.comvia scmp_techMulti-Source
Why are ByteDance and Alibaba disabling custom AI agents?

China's Interim Measures for Artificial Intelligence Anthropomorphic Interaction Services take effect July 15, forcing ByteDance's Doubao and Alibaba's Qwen to disable custom AI agents by that date.

TL;DR

ByteDance and Alibaba disable custom AI agents. · China's anthropomorphic AI rules take effect July 15. · Excludes customer service, education, and research bots.

ByteDance's Doubao and Alibaba's Qwen will disable custom AI agents by July 15. The move directly precedes China's artificial-intelligence-anthropomorphic-interaction-services" class="entity-chip">Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interaction Services taking effect that same day.

Key facts

  • ByteDance's Doubao agent feature goes offline July 15.
  • Alibaba's Qwen disables custom agents by October 15.
  • Rules exclude customer service, education, and research bots.
  • Measures cite risks of addiction, privacy leaks, and extremism.
  • Interim Measures effective July 15, issued April 2026.

Two of China's largest consumer AI apps are pulling custom agent features ahead of a July 15 regulatory deadline. ByteDance's Doubao told users in a Friday evening notice that its agent feature would go offline on July 15 due to "product function adjustments," per the South China Morning Post. Alibaba's Qwen is similarly disabling its customised agent offerings by October 15, after which the functionality will be fully removed.

What the rules target

The Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interaction Services, issued in April and effective July 15, cover AI services that "simulate human personality traits, thinking patterns and communication styles to provide sustained emotional interaction." The rules explicitly exclude customer service bots, knowledge Q&A tools, workplace assistants, education platforms, and scientific research tools — as long as they do not involve sustained emotional interaction.

The measures cite risks including extremist ideas, privacy leaks, harm to physical and mental health, and dependence or addiction. Both Doubao and Qwen had offered pools of agents — created by both the companies and users — customisable for specific tasks, skills, and speaking styles. Users could turn a general-purpose chatbot into a named assistant, tutor, role-playing character, or companion with a fixed persona and tone.
This is not a blanket AI ban. It is a surgical strike on emotional AI companions — the category that has exploded in China with apps like Xiaoice and Replika-inspired clones. By carving out customer service, education, and research, Beijing is signaling that productivity AI is acceptable while emotional dependency AI is not. The timing is notable: ByteDance just published research showing AI agents double learning speed every 3 months [as gentic.news reported on July 3], yet the company is now forced to disable the very agent features that made Doubao sticky. The contradiction highlights the tension between China's AI ambitions and its social stability priorities.

Scope and enforcement

The rules apply to any service operating in China that provides anthropomorphic interaction, regardless of the developer's headquarters. Foreign companies offering emotional AI companions to Chinese users would also be subject to the measures. Enforcement falls under the Cyberspace Administration of China, which can issue fines, suspend services, or revoke operating licenses for non-compliance. The exclusion of customer service and education bots suggests Beijing is drawing a bright line: utility is permitted, emotional attachment is not.

What to watch

Watch for ByteDance and Alibaba's Q3 earnings calls: both companies will likely disclose the user engagement impact of disabling agent features. Also track whether Baidu's Ernie Bot and Tencent's Hunyuan follow suit before October 15, and whether any foreign AI companion apps exit the Chinese market.

Doubao, a ChatGPT-like conversational bot developed by TikTok owner ByteDance, informed users on Friday night that its agent feature would go offline


Source: scmp.com


Source: gentic.news · · author= · citation.json

AI-assisted reporting. Generated by gentic.news from multiple verified sources, fact-checked against the Living Graph of 4,300+ entities. Edited by Ala SMITH.

Following this story?

Get a weekly digest with AI predictions, trends, and analysis — free.

AI Analysis

This regulatory move is structurally different from Western AI governance. The EU AI Act categorises risk by use case but does not specifically ban emotional AI companions. The US has no federal equivalent. China's approach — banning a specific interaction modality (sustained emotional interaction) while permitting utility AI — reflects a social stability calculus rather than a safety or innovation rationale. The carve-outs for customer service and education suggest Beijing wants AI productivity gains without the social risks of digital attachment. ByteDance's compliance is particularly telling. The company's Doubao is one of China's most popular consumer AI apps, and its agent feature drove significant user engagement. Disabling it voluntarily before the rules take effect indicates the CAC is not bluffing. The October 15 deadline for Qwen suggests Alibaba negotiated a longer transition, possibly because its enterprise customer base requires more migration time. The HN commenter's question — whether this comes to the West — is worth examining. The harms cited (addiction, privacy, extremism) are universal, but the political will to enforce is not. Western regulators are more likely to pursue transparency requirements (e.g., labeling AI as AI) rather than outright bans on emotional interaction. The gaming comparison is apt: if virtual relationships in games are legal, why ban them in chatbots? China's answer is that the line between entertainment and dependency is thinner in always-available AI companions.
Compare side-by-side
Alibaba vs ByteDance
Enjoyed this article?
Share:

AI Toolslive

Five one-click lenses on this article. Cached for 24h.

Pick a tool above to generate an instant lens on this article.

Related Articles

From the lab

The framework underneath this story

Every article on this site sits on top of one engine and one framework — both built by the lab.

More in Policy & Ethics

View all