The effect of AI automation and augmentation on radical innovation: from the perspective of Organizational Information Processing Theory

Drawing on Organizational Information Processing Theory (OIPT), this study examines how AI automation and AI augmentation contribute to radical innovation and how their effects vary with the information-processing requirements of innovation work. We adopt an explanatory sequential design combining a two-wave survey of 480 software engineers in manufacturing and service industries with three post-quantitative organisational case studies. The quantitative results show that both AI automation and AI augmentation are positively associated with radical innovation. However, their relative effectiveness is contingent on environmental dynamism and task connectivity: both conditions strengthen the positive relationship between AI augmentation and radical innovation while weakening the corresponding relationship for AI automation. The case evidence further explains these moderation effects by identifying two distinct information-processing mechanisms. Task connectivity creates a relational information-processing demand, under which localised automated processing becomes less sufficient as innovation increasingly requires integration across interdependent components. Environmental dynamism creates a temporal information-processing demand, under which codified routines become less durable as previously valid assumptions lose relevance over time. By conceptualising automation and augmentation as distinct human-AI information-processing arrangements, this study extends OIPT and provides a more contingent explanation of how AI supports radical innovation.