Discarding echoes of the past: A taxonomy for designing socio-technical unlearning artifacts

Unlearning outdated knowledge is crucial for organizations to adapt to environmental needs and foster innovation. While unlearning is an auspicious approach, little is known about how to design supporting tools. This paper explores how organizations can effectively implement unlearning interventions and presents an overview of design options for unlearning support systems. By combining deductive and inductive reasoning, we crafted an artifact in the form of a taxonomy. The taxonomy synthesizes insights from diverse research domains, takes a socio-technical stance, and thereby aims to bridge the theory-practice gap in implementing unlearning. To investigate the taxonomy’s applicability, we conducted evaluation sessions with experts and employed it to describe three illustrative unlearning cases. With our work, we seek to unleash the potential of unlearning by guiding the design of supporting digital tools and aligning existing insights beyond disciplinary boundaries.