From Leaders to Algorithms : Perceptions of AI Replacing Middle Managers in Swedish Conventional Organisations

Sammanfattning

Algorithmic management (AM) on gig platforms automates functions previously performed by human managers. Whether this possibility extends to middle managers in conventional organisations is rarely discussed openly. This thesis investigates how organisational actors in Swedish conventional organisations perceive the possibility that AM enabled by artificial intelligence (AI) could replace middle managers, and what keeps the question off the organisational agenda. Eighteen semi-structured interviews across twelve Swedish organisations were analysed through four sensitising concepts: Digital Taylorism, the management/leadership boundary, AI framing, and the Nordic institutional filter. The replacement question is structurally not asked. Two discursive moves work as a paired closed circle, each catching the question the other cannot answer alone. A leadership claim names motivation, trust and judgement as work AI cannot do, and a tool framing of AI presents it as helping rather than replacing managers. A measurement gap keeps both claims untestable, because management work has not been decomposed into specifiable units. The Swedish institutional setting was tested as a candidate explanation. Within the perspectives obtained, institutions shape the mechanism of change but do not block the conversation. The constraint on articulating replacement is the discursive lock, not the institutional setting. The thesis refines the literature's assumption that algorithmic management is platform-bound. The platform-conventional split tracks where the Taylorist decomposition of management work has been performed, and conventional Swedish workplaces sit on the earlier side, which is why the measurement gap holds and the discursive lock survives at middle-management level.