Veranstaltungsprogramm der VHB Jahrestagung

Decision-making is an area that is undergoing rapid changes due to the introduction of artificial intelligence (AI), as AI recommender systems can help mitigate human flaws and increase decision accuracy and efficiency. However, AI can also commit errors or suffer from algorithmic bias. Hence, blind trust in technologies carries risks, as users may fail to recognize mistakes and follow advice that could be harmful to others and to themselves. The current study investigates whether users overrely on AI advice—to their own detriment and that of other parties. In an online behavioral experiment, we identify that people follow advice more frequently when they know it to be AI-generated advice. This is particularly concerning in cases when people follow a recommendation that contradicts available context information as well as their own assessment. Frequently, this overreliance leads not only to inefficient outcomes for the advisee, but also to unjustified discrimination of others. The results call into question how AI is being used in assisted decision making, emphasizing the importance of AI literacy and effective trust calibration for productive employment of such systems.

The purpose of this study is to provide an alternative measure of value-based management (VBM) sophistication that relies on a dictionary-based natural language processing (NLP) method. First, we develop and validate a customised dictionary for VBM sophistication. Second, we apply the novel measure to explore the relation between VBM sophistication and corporate social responsibility (CSR) decoupling. The findings indicate that our novel measure has acceptable convergent and predictive validity. We also find that VBM sophistication mitigates CSR decoupling. This study first contributes a customised, open-source dictionary for the measurement of VBM sophistication. Second, it assesses the validity of the novel measure through content, convergent, and predictive validity. Third, we integrate the literature on VBM and CSR by showing that VBM curbs CSR decoupling. Overall, our study generates future research opportunities to leverage NLP and advance the integration of textual analysis into management accounting research.

This study empirically investigates the impact of implementing a performance reporting system on a critical facet of medical care quality – specifically, the timeliness of discharge letter completion following patients' hospital stays. Leveraging longitudinal archival data from three distinct departments (clinics) within a leading German tertiary medical center, we examine the dynamic effects of the reporting system on performance in three distinct phases: prior to its implementation (2014), during its active utilization (January 2015 to July 2019), and subsequent to its discontinuation (August 2019 to December 2021). Additionally, we scrutinize the influence of service complexity on the efficacy of performance reporting, employing the number of external inputs per discharge letter (e.g., laboratory tests, diagnostic procedures) as a proxy for service complexity and division of labor. Our primary findings reveal that the introduction of a performance reporting system can enhance performance across the board, irrespective of service complexity, provided it is accompanied by appropriate governance structures. Furthermore, our analysis indicates that the timely release of patient discharge letters is a process that does not naturally sustain itself. In fact, for highly efficient departments, we observe a deterioration in performance after the management opted to discontinue the reporting system.