Reactive Users vs. Social Recommender Systems: Managing Opinion Drifts with Adaptive Policies
For users of social recommender systems, this work provides a theoretical understanding of how individual-level adaptive strategies can mitigate algorithmic opinion biases.
This paper investigates whether reactive users who adaptively adjust their content consumption can limit opinion drifts induced by social recommender systems. It shows analytically and through simulations that an adaptive policy can prevent opinion drifts and, when users prioritize opinion preservation, can yield higher expected utility than a fixed policy.
Recommendation systems are used in a range of platforms to maximize user engagement through personalization, promotion of popular content, and the use of information from social networks. It has been found that such recommendations may shape users' opinions over time. In this paper, we ask whether reactive users, who are cognizant of the influence of the content they consume, can limit such changes by adaptively adjusting their content consumption choices. To this end, we study users' opinion dynamics under two stochastic content consumption policies: a passive policy, where the probability of clicking on recommended content is fixed, and a reactive policy, where the probability of content consumption adaptively decreases following large opinion drifts. We analytically derive the expected opinion and user utility under these policies when a user is influenced by both a social network and the recommender. We show that the adaptive policy can help users prevent opinion drifts induced by recommendations and that when a user prioritizes opinion preservation, the expected utility of the adaptive policy can outperform the fixed policy. We validate our theoretical findings through numerical simulations. These findings help better understand how user-level strategies can challenge the biases induced by recommendation systems.