(Yicai) July 6 -- Chinese tech giants ByteDance and Alibaba Group Holding disabled anthropomorphic artificial intelligence agent features on their AI applications due to regulatory uncertainties, leading to widespread dissatisfaction among users.

ByteDance's Doubao and Alibaba's Qwen will remove user-defined AI agent functions on July 15, the pair announced at the same time on June 3. These features are distinct from the broader concept of general AI agents and specifically refer to AI personas created by individual users, primarily focusing on emotional companionship.

In April, Chinese regulators introduced the Interim Measures for the Management of Anthropomorphic AI Interactive Services, which will take effect on July 15. The measures call for "inclusive yet prudent and tiered classification-based regulation" of relevant services, but specific regulatory standards and implementation plans have not yet been released, leaving developers uncertain about compliance pathways.

The decision by ByteDance and Alibaba to remove the features instead of addressing compliance issues from a technical standpoint has led to user unrest in China. Some netizens noted that if companies are concerned about these features affecting minors, solutions such as real-name verification or offering paid options could be implemented rather than shutting down the entire functionality.

In response to concerns about data retention after the removal of the features, ByteDance and Alibaba said that users can save their agent configuration information and important conversation records through screenshots, conversation exports, and other methods. Data would be cleared after a certain period, the pair pointed out.

ByteDance users can transition to Maoxiang, another AI interactive application under the TikTok owner, to create and use AI agent services, the company said.

The decision by ByteDance and Alibaba to remove human-like AI agent features is not a tightening of their AI business, but rather a compliance adjustment, analysts pointed out. Other firms may also gradually separate high-risk anthropomorphic interaction features from mainstream platforms, redirecting them to independent products for specialized operations to meet regulatory requirements, they added.

Editors: Dou Shicong, Martin Kadiev