<![CDATA[Automated decision-making (ADM) is becoming a significant feature of contempo- rary public administration, with major implications for how bureaucratic decisions are made and discretion exercised. Existing scholarship has mostly examined ADM within single sectors, leaving a limited cross-sector understanding of how discretion shifts across policy domains. This article addresses this gap by analyzing how empir- ical studies describe changes in bureaucratic discretion after ADM implementation across sectors. Using a systematic literature review of 21 empirical studies published between 2020 and 2025, this study discovers and examines five sectors: social wel- fare and public health, labor market governance, administrative systems, policing, and judicial decision-making. The findings show that ADM does not transform dis- cretion uniformly but reconfigures it according to the sectoral decision structures. Three analytical dimensions were identified: level of automation, level of shift, and type of shift. Highly standardized sectors, such as labor market governance, show stronger automation and a greater curtailment of discretion. The social welfare, ad- ministrative, and judicial sectors retain stronger human involvement because of their contextual complexity, accountability, and reviewability. Policing shows the strongest continuity of discretion owing to situational judgment and professional autonomy. Overall, this study contributes to ADM scholarship by offering a cross-sector perspec- tive on bureaucratic discretion and suggesting that public organizations tailor ADM design and implementation to sector-specific decision logics to preserve accountabili-ty and meaningful human judgment.]]>
