Options
Technical autonomy in human-machine collaboration: An exploration of human responses to advanced technology
Type
doctoral thesis
Date Issued
2022-09-19
Author(s)
Abstract
Based on the assumptions that technology is becoming increasingly autonomous and new human-machine constellations are emerging, this thesis explores the question of how humans respond to such increasingly autonomous technical systems. Although many tasks are accomplished in close collaboration between humans and machines, existing research insufficiently incorporates the user's (or collaborator's respectively) perspective. Three articles address the overarching research question. The first article introduces the concept of technical autonomy, derives five levels of this property, and distinguishes it from the notion of automation. For exploring human responses, two online experiments are conducted. The first experiment tests the effect of technical autonomy on general social perceptions (N = 288), and the second experiment analyzes the effect of technical autonomy in organizational decision-making scenarios (N = 341). The thesis presents evidence that the level of autonomy affects how people respond to technical systems. It suggests that people perceive technical systems that are high in autonomy as having more agency and competence, and are more willing to rely on them than systems operating on lower levels of autonomy. Third, this effect is particularly evident when technical systems are described as adaptive or both, adaptive and open. The results of this research will inform both research and practice. For the former, the results suggest that technical autonomy is a relevant variable in human-machine collaboration and this technical property should be further incorporated into research in general and theory-building. For the latter, the findings suggest that autonomous systems should be implemented with caution.
Abstract (De)
Ausgehend von den Annahmen, dass Technik immer autonomer wird und gleichzeitig neue Mensch-Maschine-Konstellationen entstehen, geht die vorliegende Thesis der Frage nach, wie Menschen auf zunehmend autonome technische Systeme reagieren. Obwohl viele Aufgaben in enger Zusammenarbeit zwischen Menschen und Maschinen erledigt werden, bezieht die bestehende Forschung die Perspektive der Nutzerin bzw. des Nutzers nur unzureichend ein. Insgesamt drei Artikel befassen sich mit dieser Forschungsfrage. Der erste Artikel führt das Konzept der technischen Autonomie ein, leitet fünf Stufen dieser Eigenschaft ab und grenzt sie vom Begriff der Automatisierung ab. Um die menschliche Sicht abzubilden, wurden danach zwei Online-Experimente durchgeführt. Das erste Experiment untersucht die Auswirkungen technischer Autonomie auf die allgemeine soziale Wahrnehmung (N = 288), und das zweite Experiment analysiert die Auswirkungen technischer Autonomie in Entscheidungssituationen (N = 341). Die Ergebnisse deuten darauf hin, dass sich sowohl die soziale Wahrnehmung als auf die Bereitschaft auf ein technisches System zurückzugreifen in Abhängigkeit des Autonomiegrades verändern. Menschen scheinen technische Systeme mit einem hohen Autonomiegrad als handlungsfähiger, kompetenter und akzeptabler einzuschätzen als Systeme mit einem niedrigeren Autonomiegrad. Dieser Effekt zeigt sich, wenn Systeme als lernfähig beziehungsweise als lernfähig und offen beschreiben werden. Für die Wissenschaft bedeuten die Ergebnisse, dass technische Autonomie eine relevante Variable in der Mensch-Maschine-Kollaboration ist und diese technische Eigenschaft weiter in die Forschung im Allgemeinen und die Theoriebildung im Besonderen einbezogen werden sollte. Für die Praxis bedeuten die Ergebnisse, dass die Implementierung autonomer Systeme umsichtig und sorgfältig erfolgen soll.
Language
English
Keywords
Mensch-Maschine-Kommunikation
Künstliche Intelligenz
Soziale Wahrnehmung
Entscheidungsunterstützung
EDIS-5218
human responses
decision support
Human-machine-collaboration
autonomous technology
taxonomy
HSG Classification
not classified
HSG Profile Area
None
Publisher
Universität St. Gallen
Publisher place
St.Gallen
Official URL
Subject(s)
Division(s)
Eprints ID
267348
File(s)