China’s attempts to regulate AI
China advances AI governance via three main initiatives. Firstly, the Cyberspace Authority of China (CAC) released in 2021a draft set of thirty rules for regulating internet recommendation algorithms such as TikTok or news apps. One of the rules requires algorithm developers to explain the functionality of algorithms in the case of infringement of user rights and interests. The algorithmic explainability is one of the main challenges for AI regulators worldwide. CAC is the most advanced and influential Chinese regulator in the field of AI.
Secondly, the China Academy of Information and Communications Technology (CAICT) tries to develop testing and certification systems for ‘trustworthy AI’. CAICT’s initiative is similar to EU and USA initiatives aiming towards ‘trustworthy AI’. However, it is weak on implementary mechanisms.
Thirdly, the Ministry of Science and Technology (MOST) initiative focuses on ethical guidelines for AI. MOST proposes a decentralised and self-governance approach by inviting tech companies to establish their own ethics review committees.
China’s AI governance experiment could contribute to global efforts to regulate the development of AI.