When the bridge is not human: Algorithmic interference in forming social relationships through the manipulation of weak ties

Abstract In recent years, social media applications have grown in number and in user bases. Recommendation algorithms on these platforms refer social others and related content to users. Using Granovetter's tie strength theory and the literature on relationship formation as conceptual foundations, we argue that these social media algorithms can damage a user's ability to establish diverse relationships and the benefits therein, thereby reducing personal, and, when aggregated, societal advantages. We argue that this occurs because the algorithms take on social actor roles and operate as “weak tie imposters” that serve as bridges to like others and content. This work provides a new conceptualization of the role recommendation algorithms play in social relationships, argues how they impact social relationship development and user privacy, and offers potential solutions to the issues related to algorithmic interference.