Public organizations have increasingly adopted artificial intelligence (AI) to inform the decision-making processes that underpin public services design and delivery (Ammitzbøll Flügge et al., 2021; Strich et al., 2021). Scholars have provided evidence for the positive outcomes generated by AI support to public decision-making, such as the rationalization of the administrative workflow and the provision of improved public services (Misuraca et al., 2020). However, recent works have also explained the unexpected outcomes generated by the adoption of AI to inform the decision-making processes of public organizations, and hence the services they provide (Medaglia et al., 2021; Wirtz et al., 2020). Following recent controversial cases of AI adoption in the public sector, civil society and public audiences have raised calls for further scrutiny of the use of AI to inform public services design and delivery (de Bruijn et al., 2021; Grimmelikhuijsen, 2023). Academic research has echoed these concerns by showing increasing attention to the impact AI has on decision-making processes in public organizations (Giest and Klievink, 2022; Lorenz et al., 2021). In line with this research, we aim to shed light on how AI logic permanently alters public organizations’ decision-making process, and how these transformations generate consequences for the accountability of public organizations (Busuioc, 2021).

To capture the nuanced effects generated by AI on accountability, it is necessary to explain their very specific characteristics that determine organizational, legal, and institutional transformations in the public decision-making processes, and hence on services design and delivery. Unlike other technological systems, AI standardization introduces a degree of opacity in the workflows it mediates (Burrell, 2016; Leonardi and Treem, 2020). AI reasoning does not allow human actors to properly understand and explain the AI functions (Zhang et al., 2021) that inform the decision-making processes (Strich et al., 2021).