Analyzing Privacy Dynamics within Groups using Gamified Auctions
Hüseyin Aydın, Onuralp Ulusoy, Ilaria Liccardi, Pınar Yolum
TL;DR
This work investigates privacy dynamics in multi-owner content using RESOLVE, an auction-based game where a participant and three software agents bid on sharing a co-owned image. By implementing a two-round group decision process with budgets, taxes, and rewards, the study reveals that individual privacy valuations shift when content becomes group-owned and that participants both defend and concede privacy depending on outcomes and budgets. The findings show a privacy paradox in action and delineate user profiles and behaviors that can inform the design of privacy management tools for collaborative platforms. Overall, RESOLVE provides empirical insights into group privacy decision-making with practical implications for developing co-owned-content privacy controls in social and collaborative environments.
Abstract
Online shared content, such as group pictures, often contains information about multiple users. Developing technical solutions to manage the privacy of such "co-owned" content is challenging because each co-owner may have different preferences. Recent technical approaches advocate group-decision mechanisms, including auctions, to decide as how best to resolve these differences. However, it is not clear if users would participate in such mechanisms and if they do, whether they would act altruistically. Understanding the privacy dynamics is crucial to develop effective mechanisms for privacy-respecting collaborative systems. Accordingly, this work develops RESOLVE, a privacy auction game to understand the sharing behavior of users in groups. Our results of users' playing the game show that i) the users' understanding of individual vs. group privacy differs significantly; ii) often users fight for their preferences even at the cost of others' privacy; and iii) at times users collaborate to fight for the privacy of others.