Building Knowledge through Action: Considerations for Machine Learning in the Workplace

Abstract

Innovations in machine learning are enabling organisational knowledge bases to be automatically generated from working people's activities. The potential for these to shift the ways in which knowledge is produced and shared raises questions about what types of knowledge might be inferred from working people's actions, how these can be used to support work, and what the broader ramifications of this might be. This paper draws on findings from studies of (i) collaborative actions, and (ii) knowledge actions, to explore how these actions might (i) inform automatically generated knowledge bases, and (ii) be better supported through technological innovation. We triangulate findings to develop a framework of actions that are performed as part of everyday work, and use this to explore how mining those actions could result in knowledge being explicitly and implicitly contributed to a knowledge base. We draw on these possibilities to highlight implications and considerations for responsible design.