From algorithmic disengagement to algorithmic activism: Charting social media users’ responses to news filtering algorithms

•Not only algorithms influence people, but also people influence algorithms.•Cluster analysis identified different types of user engagement with algorithms.•Internet skills and political efficacy affect user engagement with algorithms. This study investigated the responses of social media users encountering algorithms. Surveying 3441 Americans, four types of users were identified through latent class analysis: 1) the disengaged who don’t exhibit much conscious responses; 2) the negative curators who mainly unfollow and block connections to consume less news; 3) the positive curators who mainly follow and add connections to consume more news; 4) the activists who attempt to challenge the algorithms. It is also found that the users’ level of Internet skills and political efficacy influenced their engagement with algorithms such that those with higher Internet skills and political efficacy were more likely to be classified as activists. The findings of this study resonate with the existing digital divide research that suggest human psychology and capacity determine meaningful technology use. Overall, this study’s focus on users shifts the understanding of algorithms from a technical force to a social and personal one.