China Breaks New Ground with Draft Algorithm Regulation Rules
China's cyberspace regulator releases pioneering draft rules targeting recommendation algorithms used by online services. Experts note the regulations introduce novel measures but leave practical implementation questions unanswered.
China's Cyberspace Administration releases pioneering draft regulations that target algorithmic recommendation systems in online services. By issuing rules specifically for recommender algorithms, China breaks new regulatory ground in artificial intelligence governance. The translated provisions require providers to comply with existing laws, secure user data, and actively monitor their systems.
Some elements of the draft reflect China's controlled information environment by directing providers to uphold mainstream value orientations and disseminate positive energy. Other sections introduce brand-new measures specifically designed for recommender algorithms, though it remains unclear how these rules will apply in practice. For example, the draft prohibits apps from engrossing or addicting their users, which creates direct compliance challenges for highly engaging short video platforms like Douyin.
The regulations also aim to give users more control over their recommendations by allowing them to see, edit, and delete user tags. However, explicit tags represent only a small fraction of how complex algorithms actually function, as systems rely heavily on nuanced behavioral data like watch time and click patterns. It is entirely unclear how platforms could realistically display this underlying data to satisfy the new requirements.