Artificial intelligence in the workplace State of research, concepts and empirical evidence regarding autonomy

Abstract

The current debate about artificial intelligence (AI) often expects a radical change in the working world, which also affects the quality of work. So far, many far-reaching expectations face very little empirical evidence. Moreover, it is often not considered that the use of AI at the workplace requires certain digital work tools, which themselves can influence core aspects of work quality, such as autonomy. This paper introduces the current state of research on AI alongside approaches on possible relationships between AI and autonomy as a core dimension of work quality. We empirically investigate these relationships using the SOEP-IS 2019 dataset, with variables for autonomy, for various digital work tools, and for AI use. Our empirical results show only a superficial positive relationship between AI usage and autonomy. Our analyses confirm that the use of AI is systematically related to specific digital work tools, some of which show a significantly stronger relationship with autonomy than AI usage itself. The paper presents empirical findings on the importance of AI for autonomy and finds no evidence for radical differences at the workplaces. We conclude by discussing how empirical research and theories can be advanced in a context where parts of the debate on AI are oriented toward fictions of radical change.