TY - JOUR

T1 - Organisation und Algorithmus

T2 - Wie algorithmische Kategorien, Vergleiche und Bewertungen durch Organisationen relevant gemacht werden

AU - Büchner, Stefanie

AU - Dosdall, Henrik

N1 - Funding Information: Der vorliegende Text hat maßgeblich von den instruktiven Anmerkungen der Herausgeberinnen, Bettina Heintz und Theresa Wobbe, profitiert. Dies gilt auch für die kritischen und konstruktiven Diskussionen während des Workshops an der Humboldt Universität für das vorliegende Sonderheft. Zu danken ist ebenfalls Katharina Braunsmann, Korbinian Gall und Justus Rahn sowie den Teilnehmerinnen und Teilnehmern des Kolloquiums am Lehrstuhl für Organisations- und Verwaltungssoziologie der Universität Potsdam für wertvolle Hinweise und Anmerkungen.

PY - 2021/8

Y1 - 2021/8

N2 - This article analyzes how organizations endow algorithms, which we understand as digital formats of observation, with agency, thus rendering them actionable. Our main argument is that the relevance of digital observation formats results from how organizations embed them in their decision architectures. We demonstrate this using the example of the Austrian Public Employment Service (AMS), which introduced an algorithm in 2018 to evaluate the chances of unemployed persons being reintegrated in the labor market. In this regard, the AMS algorithm serves as an exemplary case for the current trend among public organizations to harness algorithms for distributing limited resources in a purportedly more efficient way. To reconstruct how this is achieved, we delineate how the AMS algorithm categorizes, compares, and evaluates persons. Building on this, we demonstrate how the algorithmic model is integrated into the organizational decision architecture and thereby made actionable. In conclusion, algorithmic models like the AMS algorithm also pose a challenge for organizations because they mute chances for realizing organizational learning. We substantiate this argument with regard to the role of coproduction and the absence of clear causality in the field of (re)integrating unemployed persons in the labor market.

AB - This article analyzes how organizations endow algorithms, which we understand as digital formats of observation, with agency, thus rendering them actionable. Our main argument is that the relevance of digital observation formats results from how organizations embed them in their decision architectures. We demonstrate this using the example of the Austrian Public Employment Service (AMS), which introduced an algorithm in 2018 to evaluate the chances of unemployed persons being reintegrated in the labor market. In this regard, the AMS algorithm serves as an exemplary case for the current trend among public organizations to harness algorithms for distributing limited resources in a purportedly more efficient way. To reconstruct how this is achieved, we delineate how the AMS algorithm categorizes, compares, and evaluates persons. Building on this, we demonstrate how the algorithmic model is integrated into the organizational decision architecture and thereby made actionable. In conclusion, algorithmic models like the AMS algorithm also pose a challenge for organizations because they mute chances for realizing organizational learning. We substantiate this argument with regard to the role of coproduction and the absence of clear causality in the field of (re)integrating unemployed persons in the labor market.

KW - Algorithms

KW - Digital observation formats

KW - Digitization

KW - Organizational learning

KW - Public organizations

UR - http://www.scopus.com/inward/record.url?scp=85109028462&partnerID=8YFLogxK

U2 - 10.1007/s11577-021-00752-0

DO - 10.1007/s11577-021-00752-0

M3 - Artikel

AN - SCOPUS:85109028462

VL - 73

SP - 333

EP - 357

JO - Kolner Zeitschrift fur Soziologie und Sozialpsychologie

JF - Kolner Zeitschrift fur Soziologie und Sozialpsychologie

SN - 0023-2653

ER -