Send message to Author

Paper short abstract

Focusing on processes within the epistemic order of clinical professions, I analyse how stakeholders align machine‑learned with established clinical knowledge in intensive and emergency care, thereby constructing illusive familiarity to traditional ways of producing and applying medical knowledge.

Paper long abstract

Initial studies in the medical professions paint an ambivalent picture. Machine learning algorithms are either conceived as disrupting epistemic practices (Anichini & Kotras 2024; Heinlein 2026), or as being domesticated by medical professionals (Williams et al. 2024; Lebovitz et al. 2022).

To systematically investigate the epistemic influences of machine learning-based CDSSs on physicians as professionals, I propose to examine the entire journey of medical knowledge: from its production, through its shaping and tailoring, to its mobilization in practice, to consider key processes within the epistemic order of clinical professions (Nerland 2012; Knorr-Cetina & Reichmann 2015).

A focus on these processes reveals practices to intersect and align machine learned knowledge with established knowledge practices at different junctures along the journey that machine-learned clinical knowledge takes: by producing medical knowledge through established practices of technology development, by obscuring its divergences to traditional forms of medical knowledge, and by aligning it with clinical workflows. Taken together, these practices reveal a multi-layered construction of what I call illusive familiarity.

Drawing on three cases of building and applying machine learning-based CDSSs in clinical intensive and emergency care, I outline how illusive familiarity of machine learning knowledge and machine learning-based CDSSs is constructed at different stages of the knowledge journey, focusing on frictional approaches to aligning different knowledge forms.

Ultimately, the relation of clinical professionals to machine learning-based CDSSs thus becomes functionally equivalent to knowledge claims of trusted colleagues, which differentiates them from rules-based medical technologies.