Although artificial intelligence (AI) offers numerous workplace conveniences to employees, it also poses challenges, such as occupational substitution risk and pressure to adapt to technology. Although existing studies have extensively examined the psychological stress induced by AI and its capacity for innovation, scant scholarly attention has been devoted to understanding how employees proactively manage such technological stress via job crafting—an interpersonal and cognitive process that may subsequently shape their work engagement. Grounded in the transactional model of stress (TMS) and work engagement theory, using the data collected from May to June 2025, comprising 319 valid questionnaires administered across diverse service-sector contexts, including hospitality, retail, catering, aviation, and e-commerce platforms, multiple regression analyses were employed to examine the mediating role of employees’ cognitive appraisals in linking AI-induced stress to work engagement, and evaluate the extent to which information technology (IT) mindfulness moderates this sequential mechanism. Results indicate that AI-induced stress positively enhances employees’ work engagement via challenge appraisal but negatively impairs work engagement through hindrance appraisal. Meanwhile, IT mindfulness exerts a significant moderating effect on the relationship between AI-induced stress and employees’ cognitive appraisals. In particular, employees with high IT mindfulness tend to perceive AI-related stress as a challenge, thereby boosting work engagement, whereas those with low IT mindfulness are more inclined to frame AI-induced stress as a hindrance, leading to reduced work engagement. The conclusions enhance the theoretical model explaining how technological stress influences employees’ psychological and behavioral outcomes. Furthermore, they offer valuable practical guidance for organizations to improve employees’ coping strategies for technological stress and balance efficiency with employee well-being during AI implementation.

This work is licensed under a Creative Commons Attribution 4.0 International License.