Mellan mänskligt omdöme och artificiell intelligens. En studie om chefers autonomi, ansvar och tillit i AI-assisterade beslutsprocesser

Sammanfattning

This thesis examines how autonomy, responsibility and trust are affected by managers’ use of AI-based decision support systems in organizations. The background of the study is the increasing use of AI in organizational decision-making processes, where technological systems are increasingly used for analysis, information management and the development of decision-making material. Previous research has mainly focused on AI’s impact on efficiency and productivity, while there is still limited knowledge of how managers’ use of AI affects their scope for action, distribution of responsibility and trust in practice. The study adopts a qualitative research approach and is based on semi-structured interviews with ten interview participants in managerial roles from different industries and organizations where AI is used as support in decision-making. The empirical material was analyzed thematically based on the study’s theoretical framework. The results show that AI-based decision support systems have gained an increasingly integrated role in organizational decision-making processes. The systems are perceived to contribute to more efficient information management, faster analyses and improved decision-making material, while also influencing which perspectives and alternatives become visible early in the decision-making process. The study also shows that ultimate responsibility is still perceived to remain with the human decision-maker, even though AI often has significant influence over analyses and recommendations. Trust emerges as a central factor in how AI is used, with internal AIsystems generally perceived as more reliable than public AI tools. Overall, the results show that autonomy, responsibility and trust are closely interconnected dimensions that continuously influence one another in AI-assisted decision-making processes.