China's cyberspace regulator has taken significant action by removing over 14,000 artificial intelligence (AI) products as part of its “Qinglang” cleanup campaign, which commenced in April 2026. This initiative aims to strengthen domestic regulations concerning AI technologies and ensure user safety across the internet.
Details of the Cleanup Campaign
The Cyberspace Administration of China (CAC) disclosed that more than 6 million pieces of harmful or illegal information were purged, alongside the suspension of over 26,000 accounts. Additionally, 1,300 listings related to AI products were deleted along with nine open-source datasets deemed illegal. These efforts reflect a major push to eliminate non-compliant and potentially dangerous AI services.
Compliance Requirements for AI Companies
As part of the new regulations, all companies providing AI services to Chinese users must adhere to strict registration, filtering, and labeling requirements. This framework is designed to enhance user safety by mandating robust safety filters and ensuring proper management of training data. Non-compliance with these rules may lead to penalties or the actual removal of services.
Targets of the First Phase
The initial phase of the campaign focused on four key issues:
- Companies that failed to register large models as required.
- Platforms lacking adequate safety reviews and filtering processes.
- Instances of AI data poisoning.
- Unlabeled AI-generated content.
To support these initiatives, local internet offices have also implemented self-check systems paired with routine monitoring and technical screening. Notable companies like Huawei, Alibaba, Zhipu, and DeepSeek have begun enhancing their compliance measures, including the introduction of special reviews in app stores and advanced content identification systems.
The establishment of a dedicated channel for the public to report AI-related abuses further emphasizes the government's commitment to enforcing these regulations and safeguarding the digital environment.
