Mellan människa och maskin : Hur revisionsassistenters erfarenhet formar tillit till AI-bottar och påverkar beslutsfattande.

Sammanfattning

As the Big Four increasingly invest in generative artificial intelligence (AI), AI-chatbots havebecome an integrated component of audit work, reshaping auditors’ workflows anddecision-making processes. This study examines how experience affects trust in, and use of,AI-chatbots and how this in turn affects decision-making in the context of auditing.Employing a qualitative methodology based on semi-structured interviews with auditassistants from one of the Big Four firms, the study compares two groups with different levelsof experience: a group of junior audit assistants and a group of senior audit assistants. Theanalysis is based on a theoretical framework that integrates the Technology AcceptanceModel 2 (TAM2) with the Theory of Technology Dominance (TTD). The study shows thataudit assistants’ use of AI-chabots is influenced by a complex interplay of professionalexperience, system experience, social norms and the rule based logic of the auditingprofession, rather than by any linear relationship between experience and technologyacceptance. To better capture this complexity, four components are introduced: educationalintegration, interactional competence, task alignment and reflected AI-trust. The study findsthat AI-chatbots are well suited for simpler, standardised tasks, whereas the complex core ofauditing still relies on critical and professional judgement. In this context, the use ofAI-chatbots is shaped by reflected trust rather than technological dominance.