From Risk Transfer to Risk Prevention: How Field-level Dynamics Shape Patterns of Uneven Algorithmization

Abstract The introduction of Artificial Intelligence (AI), including Machine Learning (ML) algorithms and behavioral data, is often expected to rapidly reorganize markets and social fields. This paper investigates instances of uneven algorithmization, understood as the selective adoption of AI within the same market, by focusing on field-level legitimation dynamics in a moralized market. Empirically, we study the Danish life insurance industry and ask why the field converged on the transformative use of AI for risk prevention, while abandoning experimentation with the more incremental and economically proven use of AI in pricing. Integrating the Strategic Action Fields (SAF) framework with scholarship on moralized markets, we show how the legitimacy of contested technologies is negotiated in contested markets. We find this to be a collective and politically charged process that unfolds at the field level, where technology legitimacy becomes localized and use-case specific as actors assess individual applications. Successful legitimation depends on the cultivation of legitimacy alignments across firms, customers, regulators, and broader publics. Through strategic framing that addresses performance outcomes, moral expectations, and the moral history of the market, skilled social actors craft local regimes of legitimation that shape patterns of uneven algorithmization.