The algorithmic management paradox: a structured integrative review of perceived autonomy on performance in digital workplaces
Abstract
Digital workplaces increasingly delegate managerial functions such as task allocation, scheduling, monitoring, evaluation, feedback, and sanctioning to software systems. These systems can improve coordination speed, scalability, and consistency, but they also redistribute worker discretion in uneven ways. This article examines that tension as an algorithmic management autonomy paradox. Rather than asking whether algorithmic management is simply empowering or controlling, the review asks which facet of autonomy is affected, through which psychological mechanism, and with what consequences for different types of performance. A structured integrative review was conducted across Scopus, Web of Science, PsycINFO, ABI/INFORM, and the ACM Digital Library, supported by backward and forward citation tracing. The core search covered 2015 to January 2026 and used search blocks related to algorithmic management, digital monitoring, platform work, autonomy, fairness, stress, empowerment, and performance. The screening process identified 358 records, retained 276 after de-duplication, assessed 94 full texts, and produced a final corpus of 41 sources. The synthesis used a thematic coding matrix to classify algorithmic practices, autonomy facets, psychological mediators, outcome categories, moderators, evidence type, and level of analysis. The review makes a deliberately bounded contribution. It does not claim to propose a wholly new theory of algorithmic management; instead, it develops a middle-range framework that connects five groups of algorithmic practices with three autonomy facets, key psychological mechanisms, and differentiated performance outcomes. The synthesis suggests that algorithmic management is most likely to support sustainable performance when workers retain meaningful method and decision-making autonomy, when monitoring is developmental rather than punitive, and when transparency is actionable through explanation, contestability, and human review. Six propositions and a practical governance agenda are offered for future research and organizational design.