Together with Space10 we investigated the treatment of sensitive personal information in AR applications, and created a technical prototype which allows users to mask this information by choosing different levels of abstraction.This prototype was part of a broader research project on shared privacy controls between users and companies, to foster more trusted services.Using common computer vision techniques allows users to select and adjust the abstraction of books, documents, faces, people, windows and doors, before any data leaves their phones. This process is deliberately integrated into the main interface, elevating this privacy controls from sub-menus to being an accessible main interaction. The visual integration alongside a short feedback loop are key to make these controls as frictionless as possible.Private Collection shows that treating people's data with respect does not exclude the use of intelligent AR applications in the home. By outlining the appropriate technical infrastructure we demonstrated an approach to foster more trustful relationships between people and service providers through digital interfaces.The visual design is a result of multiple considerations: What does sufficient abstraction look like form user perspective, how do people perceive abstracted entities in otherwise non-abstracted environments, appropriate processing load and energy consumption on mobile devices, etc.
