Algorithmic pedagogy: How Douyin constructs algorithmic imaginaries for content creators

Algorithms have become central organising infrastructures for participatory media platforms, structuring both content production and distribution. While users engage with these algorithmic systems in pursuit of visibility and success, the underlying mechanisms remain largely opaque. This study investigates how Douyin strategically constructs algorithmic imaginaries through the Douyin Content Assistant (DCA), an official platform resource designed to instruct content creators. Drawing on a mixed-method analysis of 725 DCA videos – combining critical discourse analysis with GSDMM topic modelling – we identify four key thematic clusters that reveal how Douyin frames its algorithmic system, content strategies, video aesthetics and platform governance. We conceptualise this deliberate platform-led instruction as algorithmic pedagogy: an assemblage of discourses and practices through which the platform not only shapes creators’ perceptions of how its algorithm operates, but also prescribes normative modes of engagement that guide creators in optimising their practices for platform-defined objectives. In doing so, Douyin produces an illusion of algorithmic fairness while subtly responsibilising creators to serve its commercial and ideological priorities. This analysis demonstrates that algorithmic platforms govern not solely through technical infrastructures, but also through the imaginative construction of algorithmic understanding as a form of soft governance.