The rise of artificial intelligence (AI) in organizational decision-making has sparked concerns about its potential to diminish human moral skills, a phenomenon referred to as moral deskilling. However, differences in how employees engage with AI remain conspicuously untheorized with regard to moral deskilling. Accordingly, this paper examines the dual processes of moral deskilling and upskilling within hybrid human-AI problem-solving frameworks. We propose that the extent and type of human involvement in these processes plays a critical role in determining whether moral competencies are eroded or enhanced. In scenarios where AI dominates problem-solving, the risk of moral deskilling increases, leading to a higher likelihood of amoral decision-making and unintended unethical outcomes. Conversely, in interactive problem-solving contexts where humans and AI collaborate closely, opportunities for moral upskilling may arise, enhancing human ethical decision-making capacity. The paper introduces the concept of Cognitive Ethical Engagement as a strategy to mitigate moral deskilling, emphasizing the importance of transparency, active human involvement, and continuous ethical reflection in AI-assisted decision-making processes.
