“the Algorithm Is like a Mercurial God” – Exploring Content Creators' Perception of Algorithmic Agency on Youtube

Algorithmic systems wield substantial influence in contemporary society, particularly in the context of digital platforms (Bunz, 2012; Burrell & Fourcade, 2021; Zuboff, 2023). Since it is mostly unknown how algorithms specifically work, content creators (CCs) as a new professional group who rely on their functionality for economic reasons, are in a constant state of sensemaking regarding the characteristics and perceived preferences of the algorithm (Duffy, 2020; Bishop 2019a; DeVito, 2021).

While studies suggest that users in general and CCs in specific do indeed perceive the algorithmic structures as agentic (Shin, 2022; Singler, 2020; Wu, Pedersen, & Salehi, 2019; Cotter 2024), we still know very little about the variations within these assessments and in which ways different social dispositions and industry factors play a role in that regard. This is a matter of concern as negotiating with algorithmic systems is a fundamental part of this profession. We therefore examine the following question: In what ways do content creators perceive the YouTube algorithm as possessing agency, attributing qualities such as autonomy, intentionality, foresight, and goal pursuit to it?

We do this by conducting a thematic analysis (Braun & Clarke, 2012) of 30 interviews with German-speaking CCs on YouTube. We find that CCs perceive agentic qualities of the YouTube algorithm and that their assessment does depend on their experience and exposure to it. Four key themes were identified: The algorithm is seen as an entity that is non-transparent and largely unpredictable (1); as dynamic, intentional, and human-like (2); it exerts a great deal of power (3) and is described as communicating through metrics thus constantly creating hierarchies and reinforcing inequalities (4). Thus, our study contributes to a more in-depth understanding of the challenges of emergent professions on digital platforms as well as shine new light on human-machine-interactions from a social inequality perspective.