Algorithm awareness as a path to user dissatisfaction and surveillance concerns.

Recently, personalization of content became integral to social media platforms, as it optimizes business operations, increases customer satisfaction, and maintains a competitive edge. For social media users, personalization is convenient as it presents content based on their interests. With the importance of social media in people’s everyday life, questions are raised about the impact of algorithms on one’s identity. Algorithm responsiveness describes the extent to which personalized content relates to one’s identity, where algorithms either support one’s identity or ignore it. While algorithms may prioritize biased content or suppress a part of one's identity, users can lack cognitive resources to determine authentic content. Since personalized content relates to one’s identity, it shapes one’s self-perception, whilst users mutually shape personalized content by engaging with it. Nevertheless, people are generally not aware of algorithms or the impact of user behavior on personalized content and do not protect their privacy. Additionally, individuals’ perceptions of algorithms are contradictory, as some personalized content is refused and some is accepted. Scholars argue that algorithm awareness can determine the perception of algorithms, as when users become aware of algorithms they either appreciate the benefits of personalization or feel threatened. With algorithm awareness, users become more concerned about surveillance and are more likely to avoid personalized content, yet some scholars report the contrary. Since there is no research analyzing the aforementioned and as there are existing contradictory findings, this study aims to uncover “to what extent does algorithm responsiveness explain the impact of algorithmic awareness on personalization concerns of Instagram users living in the Netherlands?”. In this research, algorithm responsiveness is a mediator between algorithm awareness and the dependent variables surveillance concern and personalized content avoidance, while time spent on Instagram and age are control variables. By conducting a survey experiment (N=107), the level of participants’ algorithm awareness is manipulated, where explanations of algorithms are provided to two groups. However, the level of algorithm awareness did not have significant effects on algorithm responsiveness. Mediation analysis revealed that 8 out of 10 hypotheses are rejected, where algorithm responsiveness is not a significant mediator. Based on manipulated and non-manipulated algorithm awareness, perceived algorithm insensitivity increased the likelihood of personalized content avoidance. It was found that younger Instagram users perceive more responsiveness and less insensitivity and that they are more likely to avoid personalized content than older Instagram users. Based on the non-manipulated algorithm awareness, algorithm awareness increased perceived algorithm responsiveness. Thus, media entities should ensure that especially young users perceive algorithm responsiveness. Additionally, future research should utilize different stimuli and the scale for algorithm awareness and ensure a large and representative sample. The insignificant findings can be useful for scholars and policymakers, who place high importance on the level of algorithm awareness but should analyze other factors like age.