Social impact of recommendation algorithm in crisis: Forming algorithmic experience through group information interaction and algorithm task fit

This study explores the societal impact of recommendation algorithms during crisis situations, specifically examining the dynamic interaction between users, algorithms, and tasks in disaster contexts. By integrating the Task-Technology Fit (TTF) model, Stress and Coping theory, and Social Identity theory, the research constructs a comprehensive analytical framework to better understand group information behavior and algorithmic experiences. Addressing the theoretical limitations of existing research that separates human-algorithm interaction from human-human interaction, this study innovatively incorporates algorithmic performance and interaction purpose into a unified analysis model. Through an experimental design, the study manipulates the "personalized content recommendation" feature by enabling and disabling it to observe how different algorithm configurations influence user perceptions and behaviors. The findings reveal that a strong task-technology fit enhances group information interaction intentions, with personalized content recommendations playing a dual moderating role. They not only bridge perceived disaster threats and task-technology fit but also impact the relationship between perceived disaster threat, perceived interactive support, and overall algorithm experience. This study contributes to the theoretical expansion of task-technology fit applications in disaster contexts and provides practical insights for designing recommendation algorithms in crisis situations. It highlights the importance of algorithmic forming and contextual fitness in improving public engagement and crisis response efficiency.