Abstract: Based on the APCO (Antecedents-Privacy Concerns-Outcomes) model framework, this study adopts a mixed-method approach combining qualitative and quantitative research to examine the impact of users’ algorithm awareness and privacy concerns on their privacy risk coping behaviors, from the perspective of algorithm abuse. In the qualitative study, in-depth interviews with 24 social media users were conducted and analyzed with grounded theory, aiming to identify key dimensions of privacy concerns under the lens of algorithm abuse. In the quantitative study, survey data from 513 users were empirically analyzed using structural equation modeling to examine the direct effect of algorithm awareness on privacy risk coping behaviors and to further explore the mediating roles of five identified privacy concern factors. Results from the qualitative analysis reveal five key privacy concern dimensions: perceived privacy intrusion, perceived algorithm surveillance, perceived algorithm bias, perceived algorithmic decision-making risk, and perceived data permanence risk. Empirical results show that algorithm awareness has a significant positive impact on privacy risk coping behaviors. Moreover, perceived privacy intrusion, algorithm bias, algorithmic decision-making risk, and data permanence risk partially mediate this relationship, while the mediating role of perceived algorithm surveillance is not significant. These findings provide practical evidence to support the governance of algorithm abuse risks and the management of user privacy risks in China.

Key words: Algorithm abuse, Algorithm awareness, Privacy concern, Privacy risk, Coping behavior, Social media