Optimizing human–AI collaboration: the critical role and mechanism of digital literacy in hybrid intelligence workflows

Purpose This study originates from observations and reflections on a fundamental work paradigm shift: the deep integration of artificial intelligence (AI) is driving work beyond task automation toward hybrid intelligence (HI). This reveals a profound paradigm misalignment. Traditional literacy models fail to capture the relational, generative and systemic nature of HI and lack a micro-level perspective on human–AI collaboration. This study aimed to identify and explicate digital literacies and their mechanisms within HI workflows, providing both theoretical foundations and practical architecture for a more integrated, adaptive and co-evolutionary paradigm. Design/methodology/approach An exploratory qualitative design was employed, drawing on semi-structured interviews with 24 HI practitioners in China's financial and e-commerce sectors, contexts characterized by high complexity, ambiguity and uncertainty. The data were analyzed thematically to identify key literacy practices. Findings Four core literacies were identified as forming a generative HI system: cross-border literacy (integrating heterogeneous cognitive frameworks), cognitive literacy (orchestrating generative dialogue with AI), evolve literacy (enabling bidirectional co-learning) and transformative literacy (driving architecture reconstruction). These literacies function as micro-level rules addressing challenges across HI workflow stages. Originality/value This study identifies four critical literacies that catalyze HI workflows, advancing digital literacy research toward a dynamic practice-relational paradigm. It contributes a micro-process perspective to socio-technical systems theory and offers a generative account of HI as a complex adaptive system. This study provides an actionable framework elevating digital literacy from an individual trait to an organizational design cornerstone.