Reduced workload or diminished capabilities? The double-edged sword effect of artificial intelligence on knowledge workers' innovative behaviours

Purpose The rapid diffusion of artificial intelligence (AI) assistants into workplace processes has sparked an ongoing debate about whether these technologies stimulate or suppress human innovation. This study investigates the double-edged effects of AI assistants on knowledge workers' innovative behaviour, by identifying both the positive mechanisms that enhance innovation and the negative pathways that hinder it, in order to help organisations maximise the benefits of AI while mitigating its drawbacks. Design/methodology/approach Grounded in self-determination theory, this study collected questionnaire data from 400 employees across seven Chinese enterprises through situational experiment methodology to empirically investigate the double-edged sword effect of AI assistants on innovative behaviour. Findings Firstly, AI-assistant intelligence can directly enhance the innovative job performance of knowledge workers, and indirectly boost innovation by reducing perceived work intensity, particularly in organizations with high organisational AI readiness. Simultaneously, highly intelligent AI assistants may trigger personal development concerns in knowledge workers, thereby inhibiting innovative behaviour. This negative effect is also amplified when organisational AI readiness is high. Finally, the coexistence of these two opposite mechanisms confirms the double-edged sword effect. Originality/value This study leverages self-determination theory to demonstrate how AI assistants can both enhance and undermine knowledge workers' innovative performance, and highlights organisational AI readiness as a critical boundary condition, thereby contributing to the “AI-driven innovation” debate.