Leadership Behavior in the Age of AI: Comparing the Detection of Exploitative Leadership via ChatGPT with Leaders and Employee Ratings

This study examines and compares the assessment of exploitative leadership in written email communication by human leaders and employee and a customised and trained ChatGPT-based model. To evaluate perceived level of exploitative leadership, leaders and employees will be presented with a series of standardised email scenarios and asked to rate each scenario along the established exploitative leadership scale (based on Schmid et al., 2019). These leaders and employee evaluations will then be directly compared to the outputs generated by a fine-tuned ChatGPT model, which has been specifically instructed and calibrated to analyse exploitative leadership textual content.