Technical neutrality to be followed when setting up an online platform in China
❌Operators of online platforms or applications are prohibited from misusing big data, algorithms, or other technological means to engage in the following practices:
- Designing operations that go against users' will or deprive them of their right to choose;
- Influencing user choices by increasing operational complexity, disrupting continuity of use, or using similar tactics;
- Engaging in traffic hijacking, interference, or malicious incompatibility that disrupts or undermines the normal functioning of other platforms' products or services;
- Downloading, installing, or running other applications without user consent;
- Failing to provide an uninstall function for non-essential applications running on a user’s device, or creating unreasonable obstacles to uninstallation.
🚩If an online platform or app provides algorithm-based personalized recommendation service, the operator shall fulfill the following obligations:
- Clearly inform users in a prominent manner that algorithm-based personalized recommendation services are being provided, and publicly disclose the basic principles, intended purposes, and primary operational mechanisms of the algorithms in an appropriate manner.
- When using automated decision-making methods to push information or deliver commercial marketing to individuals, the platform or app shall simultaneously provide options that are not based on the individual's personal characteristics, or offer a convenient way to turn off personalized recommendation features.
- The platform or app shall not push information to minors that may induce them to imitate unsafe behaviors or acts contrary to social ethics, or that may encourage bad habits or otherwise adversely affect minors' physical and mental health. Nor shall they use algorithm-based personalized recommendation services to entice minors into internet addiction.
🔔If an online platform or application provides generative artificial intelligence service, the operator shall comply with the following requirements:
- If the platform provides generative AI services on its own, the operator shall conduct a large-language-model filing to local cyberspace administration authority in advance.
- If the platform integrates third-party generative AI services, it shall register online with local cyberspace administration authority, and shall disclose in the platform's relevant public documents the information about the LLM vendor and its contact details for users' reference.
- Online platforms providing generative AI services shall offer users a function for labeling generated or synthesized content, and shall remind users to carefully read and understand the relevant AI labeling management requirements.
- If content generated by the platform through AI is likely to cause public confusion or misidentification, the platform shall place prominent labels at reasonable positions or areas of the generated, synthesized, or edited information content to notify the public that such content originates from AI.
- With respect to content produced by generative AI, the platform shall establish and improve complaint and reporting mechanisms, set up convenient channels for submitting complaints and reports, publish the handling procedures and response timelines, and promptly accept and address public complaints and reports while providing feedback on the handling results.