Human–machine collaboration and algorithmic decision-making: The role of organizational incentives and AI capabilities

Despite growing investment in artificial intelligence (AI), organizations have yet to realize its potential for complex decision-making. We draw on the absorptive capacity (ACAP) view to conceptualize AI systems as organizational knowledge structures and incentive policies as knowledge transfer mechanisms, examining their effects on human–machine collaboration over time. We conduct a randomized controlled experiment wherein financially incentivized participants evaluate 100 loan applications with AI assistance over 10 decision rounds. Results show AI predictive power and incentive alignment influence organizational outcomes not only independently but also interactively: incentive alignment matters most when AI predictive power is low, suggesting aligned incentives may buffer organizations against the negative effects of AI degradation. Crucially, human–AI joint performance improves as experience accumulates. Further analyses suggest that cognitive capabilities, AI trust, algorithm integration, and decision quality shape individual engagement with organizational ACAP processes. Theoretically, by conceptualizing AI systems as organizational knowledge structures bridging human and algorithmic intelligence through the ACAP development processes, we argue that human–machine collaboration could serve as a new source of learning for organizations.