Generative AI Use at Work and Employee Innovative Performance: The Mediating Role of Generative AI Overreliance and the Moderating Role of AI Literacy

Generative AI (GenAI) is changing how employees work withintelligent systems, yet it is still unclear whether this leads to real innovation oronly to faster output. We draw on Job Demands-Resources (JD-R) theory tobuild and test a moderated-mediation model that links GenAI use at work,overreliance on GenAI, AI literacy, and employee innovative performance.Survey data from 246 employees of a diversified corporate group, where GenAIuse is high but AI literacy varies widely, were analyzed with PLS-SEM. Ratherthan stopping at the familiar creativity and productivity findings, we considerthe full innovative work behavior process: generating ideas, promoting them,and putting them into practice. GenAI use turns out to be a strong driver ofinnovative performance, and it also raises overreliance, while AI literacy lowersoverreliance directly and significantly. The result we did not expect is thatoverreliance does not hurt innovative performance; the harmful mediation wehypothesized simply does not appear. Put differently, heavy GenAI use neednot come at the cost of innovation as long as literacy keeps reliance undercontrol. The study separates use from overreliance, tells innovation apart fromefficiency, and shows that AI literacy works better as a direct antecedent ofhealthier reliance than as a moderator. For AI-enabled workplaces the messageis practical: encourage useful GenAI use, but invest in AI literacy at the sametime.