Introduction to the Minitrack on Design and Appropriation of Knowledge and AI Systems

The objective of this minitrack is to contribute to the body of knowledge that helps scholars and practitioners increase their collective understanding of (1) how knowledge and AI systems are planned, designed, built, implemented, used, evaluated, supported, upgraded, and evolved; (2) how knowledge and AI systems impact the context in which they are embedded; and (3) the human behaviors reflected within and induced through both (1) and (2). By knowledge and AI system, we mean a system in which human participants and/or machines perform work (processes and activities) related to the creation, retention, transfer and/or application of knowledge using information, technology, and other resources to produce informational products and/or services for internal or external customers. It is the 8th year of the minitrack. We received four papers this year and after a rigorous review process, we accepted two papers for publication in the proceedings and presentation at the conference. The first paper, co-authored by Mike Seymour, Lingyao Yuan, Alan Dennis, and Kai Riemer, examines how humans interact and perceive digital, virtual agents controlled by artificial intelligence (AI) in comparison to their interactions with real humans. The authors use a controlled lab study to show that participants rate a video human agent more trustworthy and have more affinity towards him compared to its avatar version. Further tests show that users who believe the avatar is a virtual agent controlled by AI has the same level of affinity, trustworthiness, and preferences towards the agent as those who believe it is controlled by a human. Thus, the use of a realistic digital avatar lowers affinity, trustworthiness, and preferences, but how the avatar is controlled (by human or machine) has no effect. The conclusion is that improved visual fidelity alone makes a significant positive difference and that users are not averse to advanced AI simulating human presence, whereby some may even be anticipating the use of such an advanced technology. The second paper, co-authored by Olivia Hornung, Nora Fteimi, and Stefan Smolnik, considers knowledge, as context-specific and bound to individuals, which in turn, is strongly related to human emotions such as joy or fear. More specifically, the paper adopts a holistic view to emotions across a sentiment gap analysis. Based on general sentiment dictionaries, the authors develop a dictionary aligned with KM, and subsequently apply it to KM publications to determine the presence of positive and negative emotions and categorize them according to an emotion scale. The results reveal that a variety of emotions are expressed in KM studies, both positive and negative, proving its relevance for this domain. This study also attempts to address the high diversity of terms used in regards to various emotion categories by proposing a certain degree of consolidation. Additional research opportunities related to the role of emotions in KM are also identified. We wish to thank all of the authors who submitted work for consideration in this minitrack. We also thank the dedicated reviewers for the time and effort they invested in reviewing the papers. We believe that the accepted papers contribute to furthering our understanding on the creation and appropriation of knowledge systems. We look forward to discuss these further during our session in January 2020. Proceedings of the 53rd Hawaii International Conference on System Sciences | 2020