How Humans and Machines Make Sense Together: Characteristics and Outcomes of a Case Survey Approach

Abstract

There is a growing realization of Artificial Intelligence (AI)’s importance, including its ability to provide competitive advantage and change work for the better. Indeed, organizations are investing in various AI applications in the hope to automate or augment human judgment. Despite the promise of AI, many organizations’ efforts with it are falling short. Therefore, adopting the sensemaking theory as a theoretical lens, this study investigates under which conditions and how human and machine collaborations should be structured, to enhance each other’s capabilities and facilitate optimal strategical decision-making and operational effectiveness. A framework is developed based on the level of complexity of the context and the severity of wrong decisions and four types of human-machine sensemaking processes are proposed. The framework is validated through a qualitative meta-analysis of 48 case studies of AI and highlights the characteristics of the interaction process as well as its outcomes. Besides providing a new instrument for the analysis and assessment of human-AI interactions and controls, this research aids the development of guidelines and facilitates the move towards explainable AI (XAI) design, development, and practices.