Collaborative artificial intelligence literacy and employee performance: Task–technology fit and technostress in a moderated mediation model

Abstract

Purpose: Organisations increasingly invest in collaborative artificial intelligence (AI) literacy to improve employee performance, yet many simultaneously face rising technostress. This creates a critical managerial dilemma: when do AI capability investments generate performance gains, and when do they erode the perceived fit between employees and AI-enabled work systems? This study examines how collaborative AI literacy translates into employee performance under varying levels of technostress.
Design/methodology/approach: Using a two-wave, multi-source survey of 403 employee–supervisor dyads. This study tests a conditional process model in which task–technology fit (TTF) explains the performance effects of collaborative AI literacy, while technostress acts as a boundary condition. Structural equation modelling with bootstrapping is employed to assess direct, indirect and interaction effects.
Findings/results: Results show that collaborative AI literacy improves both technology-enabled and creative performance primarily by strengthening TTF. However, technostress significantly weakens this relationship. Under high technostress, increases in AI literacy no longer enhance TTF and may even undermine it, resulting in diminished or reversed performance benefits. These findings reveal that AI literacy investments yield positive returns only when technostress is effectively managed.
Practical implications: Managers should recognise that AI literacy investments are contingent upon the prevailing stress environment. Organisations are advised to redesign training programmes to incorporate technostress resilience, prioritise task-technology alignment during AI system implementation, and treat technostress mitigation as a managerial performance metric. Such an integrated strategy ensures that AI capability development translates into sustainable improvements in operational efficiency and employees’ performance.
Originality/value: This study advances decision-oriented research by identifying a critical trade-off in AI capability investments: more AI literacy is not universally beneficial. By demonstrating that technostress can disrupt – and potentially reverse – the pathway from AI capability to performance, the study provides actionable guidance for managers on when to scale AI training and when to prioritise stress reduction and system simplification.