Automation vs. Augmentation: AI’s Dual-Edged Effects on Task Performance

Contrary to the notion that augmentation always empowers and automation invariably harms, research has rarely tested their dual-edged effects on employee outcomes. Using the Job Demands–Resources model and two-wave survey data, this study shows that automation AI use increases cognitive flexibility yet heightens AI threat awareness, whereas augmentation AI use lowers AI threat awareness but constrains cognitive flexibility. In turn, AI threat awareness undermines task performance, while cognitive flexibility enhances it, revealing two opposing psychological pathways from AI use to performance. Perceived AI knowledge operates as a boundary condition: it amplifies automation’s benefits (greater cognitive flexibility), buffers its costs (lower AI threat awareness), and eases augmentation’s adverse impact on cognitive flexibility. Its mitigation of the augmentation’s effect on AI threat awareness is small and not statistically significant. Overall, our findings move beyond “help versus harm,” offering a more nuanced, context-sensitive account of how, when, and for whom AI use facilitates, or hinders, task performance.