Abstract: Target user visibility is the basis of targeting advertising operation. The user profiles used and formed in the process of finding and tracking target users are not all identifiable data, and depicting user profiles with privacy computing technology is a process of labeling rather than identification. Tagged user profiles firstly require the data to be stealthed. In addition, the depiction of user profiles is not completed by one party subject using local data, and the formation of profiles is the result of multi-party data collaborative processing. There are two kinds of technologies for multi-party collaborative processing based on privacy computing: one is centralized computing mode, and the other is distributed computing mode. Different types and risk levels of personal information processed by different privacy computing technologies have different legitimacy conditions in different scenarios. The task of law is to establish corresponding standards.

Key words: privacy preserving computing, targeting advertising, personal information, law