The impact of generative artificial intelligence usage on job performance through job crafting and work engagement: Does digital competence matter?
Citation. Vuong, B. N. (2026). The impact of generative artificial intelligence usage on job performance through job crafting and work engagement: Does digital competence matter? Computers in Human Behavior, 178, 108921. https://doi.org/10.1016/j.chb.2026.108921
Summary. Vuong studies GenAI usage as a JD-R job resource in a survey of 758 employees in Vietnamese small and medium-sized enterprises. Using PLS-SEM, the paper finds positive associations between GenAI usage and job performance, with job crafting and work engagement acting as partial mediators. Digital competence strengthens the links from GenAI usage to job crafting and work engagement, but not the direct link from GenAI usage to job performance.
Key claims
- GenAI usage had a positive direct association with job performance (β = 0.174, p = 0.000), job crafting (β = 0.397, p = 0.000), and work engagement (β = 0.478, p = 0.000) in the PLS-SEM model.
- Job crafting (β = 0.355, p = 0.000) and work engagement (reported as β = 0.368 in Table 5 and β = 0.291 in the surrounding prose, p = 0.000) were both positively associated with job performance.
- Job crafting partially mediated the GenAI usage–job performance relationship (indirect β = 0.141, p = 0.000), and work engagement also partially mediated that relationship (indirect β = 0.176, p = 0.000).
- Digital competence moderated the relationship between GenAI usage and job crafting (interaction β = 0.074, p = 0.000) and between GenAI usage and work engagement (interaction β = 0.050, p = 0.000), but did not significantly moderate the relationship between GenAI usage and job performance (interaction β = 0.010, p = 0.208).
- The model treats GenAI usage as a job resource that may support proactive work reshaping and motivational energy, rather than only as a productivity tool or adoption outcome.
Evidence & limitations
- Evidence comes from a cross-sectional employee survey in Vietnamese SMEs; the source is useful for construct mapping and nomological-network building, but it cannot establish causal effects of GenAI usage on performance, job crafting, or engagement.
- The model uses self-report survey measures and PLS-SEM; common-method variance, occupational mix, and contextual specificity to Vietnamese SMEs limit broad generalization.
- The article reports strong explained variance (R² = 0.804 for job performance, 0.711 for job crafting, and 0.778 for work engagement), which should be treated cautiously given the cross-sectional self-report design.
- The paper contains an apparent inconsistency in the reported work engagement → job performance coefficient: Table 5 reports 0.368, while the prose reports 0.291.
- The Drive PDF includes a copyright notice and was treated as a source copy for scholarship; canonical provenance is the public DOI rather than the private Drive link.