Navigating Algorithmic Imaginaries: How Social Media Algorithms Influence Content of Science Communication

This article analyses how recommendation algorithms affect the content of science communication in social media. In the age of the anthropocene, the successful dissemination of scientific knowledge via digital media is seen as a key challenge for tackling social and environmental problems. As it is opaque recommendation algorithms that decide on the visibility of content and thereby the accessibility of knowledge on social media, winning the favor of the technology is seen as a major challenge for practitioners. Based on ethnographic fieldwork on the production of a social science format on YouTube - a joint project between a German public broadcaster and social science academics - we investigate how social media experts, who act as strategic consultants in the project, make sense of opaque social media algorithms. In particular, we examine how they translate their "algorithmic imaginaries" into practical guidelines that affect content creation practices and, consequently, the content of science communication. The practical guidelines were introduced with the promise of increasing content's visibility on the platform as they were supposedly aligned with the functions of the algorithmic system. Thus, we show that it is not the recommendation algorithms as such that influence how science communication content is presented on social media, but the sense-making surrounding the technology and the adjustment of practices towards these interpretations. Exploring the actual practices surrounding algorithmic systems is crucial to understanding the social power of such technologies. By revealing the indirect influence of algorithmic systems on the content creation practices of science communication, we unveil the hidden effects of digitalisation on the public communication of socially relevant scientific knowledge.