What question did this study set out to answer?

This research investigates how perceptions of algorithmic power and technicality influence users' resistance to algorithms on social media.

June 26, 2026Open Access

Why Users Rebel Against Algorithms: The Impact of Perceived Algorithmic Power on Fairness Evaluations, Negative Emotions, and Resistance Behaviors

YSYangyang ShiXiamen University YWYì WángBGI Group (China)JCJ Y ChenXi'an University of Science and Technology

Key Points

This research investigates how perceptions of algorithmic power and technicality influence users' resistance to algorithms on social media.
Developed a parallel mediation model based on critical theory of technology.
Examined survey data from users of Chinese algorithm-driven social media platforms.
Analyzed associations between perceived algorithmic traits, fairness evaluations, and negative emotions.
Higher perceived algorithmic power leads to lower fairness evaluations and stronger negative emotions, increasing algorithmic resistance.
Algorithmic technicality similarly fuels negative emotions and reduces perceptions of fairness, contributing to resistance.
Findings indicate that user resistance reflects broader concerns about algorithmic control beyond just recommendation accuracy.

Abstract

Platform algorithms are widely used to personalize content and organize users’ everyday social media experiences. Yet they may also become objects of resistance when algorithmic recommendations are perceived as intrusive, repetitive, or difficult to escape. Drawing on the critical theory of technology, this study develops a parallel mediation model to explain why users resist algorithm-driven social media platforms. Focusing on algorithmic power and algorithmic technicality as two perceived characteristics of platform algorithms, the model examines whether these perceptions are associated with algorithmic resistance through fairness evaluations and negative emotions. Based on survey data from users of Chinese algorithm-driven social media platforms, the results show that both algorithmic power and algorithmic technicality are associated with stronger algorithmic resistance through lower fairness evaluations and stronger negative emotions. These findings suggest that algorithmic resistance is not merely a response to inaccurate or opaque recommendations, but also reflects users’ reactions to algorithms experienced as systems of platform control and data-driven inference. By identifying fairness evaluations and negative emotions as parallel cognitive and affective pathways, this study shifts attention from algorithmic acceptance to algorithmic resistance and provides a more critical understanding of user agency in human–algorithm relations.

KI fragen

Bookmark

View Full Paper