Would you rather work with ChatGPT or a human coworker? Exploring the impact of generative AI on job satisfaction

Abstract Organizations are increasingly integrating generative artificial intelligence (AI) tools like ChatGPT for task-support purposes into the workplace. While such tools have been shown to boost productivity, it is critical to understand their psychological impact on employees’ job satisfaction if they are to be implemented effectively. Based on social response theory and the work design model, this study examines how the organizational integration of ChatGPT influences job satisfaction among employees, mediated by perceived organizational support, opportunity to perform, and job insecurity. Using a pre- and post-measurement design, we conducted a 2 × 2 between-subjects vignette experiment with a sample of 202 participants, manipulating the type of task support (i.e., ChatGPT vs. human coworker) (T1) and the nature of supervisor feedback (i.e., positive vs. negative) (T2) in the context of a corporate communication task (i.e., writing an article for the organization’s website about a strategic shift). Our results indicate that ChatGPT task support, compared to task support by a human coworker, leads to lower perceptions of both organizational support and opportunity to perform and higher perceptions of job insecurity, ultimately hindering job satisfaction. The findings suggest that, although generative AI can serve as a form of task support, its use can negatively impact employees’ perceptions. Accordingly, task support provided by human coworkers appears to still be important for maintaining job satisfaction among employees, underscoring the need for the thoughtful integration of generative AI into the workplace.