Chapter 5: ‘Real-ising’ the Benefits of Responsible Artificial Intelligence (AI) Management: An Interpretive Study

Organisational decision-makers need to manage the ever-evolving frontiers (e.g. autonomy, learning, inscrutability) of artificial intelligence (AI) responsibly, in order to navigate ethical challenges and realise organisational goals. Yet, much of the extant research on AI appears siloed and spans multiple disciplines, making it challenging to learn about management at the frontiers of AI that can support complex decision-making. The evolving expectations regarding managing AI necessitate theoretical advancements in understanding well-developed, evidence-based practices. This study approaches this challenge by situating responsible AI management practices within the design and governance of AI through a sociotechnical lens. Drawing on qualitative data, this study develops an empirically grounded model for managing AI through an interpretive approach. The model (1) delineates multifaceted responsibility (i.e. evidence, epistemic, outcome) related to AI design of and identifies associated design tactics for each dimension; (2) specifies AI governance mechanisms (i.e. structural, procedural, relational) for enacting responsible AI management practices; and (3) identifies organisational performance outcomes (i.e. instrumental, humanistic) of management practices. The research findings contribute to AI ethics, governance, and management literature, offering researchers and practitioners an empirical exposition of managing AI with actionable guidance. By turning to theory for a guide, researchers can approach managerial issues concerning future frontiers of AI with theoretical foundations, and practitioners can plan responsible initiatives that effectively manage emerging ethics challenges.