Everyday Augmented Reality (AR) headsets pose significant privacy risks, potentially allowing prolonged sensitive data collection of both users and bystanders (e.g. members of the public). While users control data access through permissions, current AR systems inherit smartphone permission prompts, which may be less appropriate for all-day AR. This constrains informed choices and risks over-privileged access to sensors. We propose (N=20) a novel AR permission control system that allows better-informed privacy decisions and evaluate it using five mock application contexts. Our system's novelty lies in enabling users to experience the varying impacts of permission levels on not only a) privacy, but also b) application functionality. This empowers users to better understand what data an application depends on and how its functionalities are impacted by limiting said data. Participants found that our method allows for making better informed privacy decisions, and deemed it more transparent and trustworthy than state-of-the-art AR and smartphone permission systems taken from Android and iOS. Our results offer insights into new and necessary AR permission systems, improving user understanding and control over data access.
https://doi.org/10.1145/3613904.3642668
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2024.acm.org/)