Sharing Same Elements in User Viewing History Data Securely Through Private Set Intersection Under User-centric Data Control

Recently, various over-the-top (OTT) streaming services as well as traditional broadcasts distribute numerous content every day, allowing users to watch their favorite content at any time. While users can choose from a quantity of content, they often view the same content mainly because most OTT streaming services implement recommendation systems. However, it is often difficult for users to realize the same content they viewed and enjoy sharing their opinions or feelings because they do not know each other when and which content they viewed. For the purpose of encouraging such enjoyable experiences, we propose a system architecture that allows users to share the same content they viewed by using the user's viewing history data. Such viewing history data is currently collected and stored by hundreds of different services and companies. Therefore, our system architecture adopts a user-centric data control model that allows users to collect and store their data on their own online storages, and use it for their purposes. If users are asked to disclose all of the raw data of their viewing history to each other or to third party when sharing, most of them will feel anxiety because the data often contains sensitive personal information. Therefore, we introduce a method using private set intersection (PSI), a cryptographic technique that allows users to share the same elements in the users’ viewing history data without revealing anything to each other except the elements at the intersection. We also demonstrate the feasibility of the architecture through use cases.