FAR054 – Enhancing Auditors’ Judgment and Decision-Making in the Age of Automation: A Learning Perspective

What?

Recent developments in technology, such as automated decision-support systems and generative AI, promise to have a significant impact on auditors’ judgment and decision-making processes
(e.g., EY, 2023; IAASB, 2021; Iacone, 2023). Emerging technologies have the potential to complete routine tasks more efficiently than staff, generating outputs to be assessed by auditors so that they will complete fewer of these tasks themselves (e.g., Susskind & Susskind, 2016; Saunders, Keune & Hawkins, 2023; Sutton, Arnold & Holt, 2023). In the short-term, this exposes auditors to two established behavioral risks: under-reliance on technology, termed algorithm aversion (e.g., Commerford, Dennis, Joe & Ulla, 2022; Commerford et al., 2024), and over-reliance on technology, known as algorithm appreciation and automation bias (e.g., Parasuraman & Manzey, 2010; Peters, 2025). Building on the Theory of Technological Dominance (Sutton, Arnold & Holt, 2018; 2023), the first aim of our project is to clarify under which conditions auditors are predicted to under- or over-rely on automated tools in the judgment and decision-making process. We try to fulfil this aim using an interdisciplinary systematic literature view, as well as interviews with auditors of all ranks and relevant members of L&D departments, national offices, and firm taskforces on technology use.

Why?

To date, performing routine tasks repeatedly under the guidance of more experienced colleagues has formed the cornerstone of staff and seniors’ on-the-job learning, enabling them to develop
necessary expertise over time through direct experiences (e.g., Westermann et al., 2015; Dierynck, Kadous & Peters, 2024; Grohnert, Gijselaers, Meuwissen & Trotman, 2025). In the medium term, as staff and senior auditors are less likely to rely on this type of learning, they may therefore miss the opportunity to develop the knowledge content and structures necessary for performing more complex judgments (e.g., Westermann, Bedard & Earley, 2015; Zhang, Thomas & Vasarhelyi, 2022; Dierynck, Kadous & Peters, 2024). Yet, such expertise remains essential for verifying the outputs of automated systems, requiring audit firms to rethink how they train junior auditors on the job, balancing efficiency in task performance with effectiveness in professional development (e.g., Mollick, 2024; Susskind & Susskind, 2016; Sutton, Arnold & Holt, 2018; 2023).
The second aim of this project is therefore to explore how on-the-job learning of staff and senior auditors can contribute to professional development and judgment and decision-making when
using automated tools. Insights from learning sciences consistently emphasize the critical role of experimentation, feedback, and reflection in expertise development, as highlighted in experiential learning theory (e.g., Kolb & Kolb, 2009). Specifically, we propose two interventions to help auditors manage contradictory/imperfect recommendations generated by automated tools. First, we propose that perceptual learning, where individuals develop knowledge implicitly through guided trials for enhanced pattern recognition, can help auditors to optimize the use of their cognitive resources (e.g., Lu & Dosher, 2022), improving both judgment and decision-making and professional development with the use of automated tools. Second, we aim to test a generative AI-based coaching tool that integrates experimentation, feedback, and reflection in young professionals’ daily work to foster novice auditors’ learning (e.g., Mollick, 2024).