Now showing 1 - 1 of 1
  • Publication
    Herausforderungen des maschinellen Lernens in der praktischen Prüfarbeit
    (Universität St. Gallen, 2023-02-20)
    Although machine learning is being applied broadly, it plays only a minor role in the conduct of audits of financial statements. The technology is thought to have a great potential to make the audit more insightful and more efficient. Nevertheless, there is a lack of systematic research into how the audit process and the application of machine learning could be combined as well as the associated challenges. This research closes this gap. By providing two separate field studies, using a constructive research approach, it shows how the audit process and the data analytics process need to be amended, to retrieve the desired results. Moreover, it demonstrates the challenges faced by auditors and how to deal with them. Two fundamentally different use cases are identified: one relying on exogenous labels (dependend variables), that allows for externally based predictions, and another based on endogenous labels, that is used for internally based predictions. Both impact the preparation of an audit, its performance and its completion.