From machine knowledge and knowledge management to a unified human-machine theory: proposal for conceptual model

Academic research in Knowledge and knowledge management tends to focus on issues related to producing, storing, organizing, sharing, and retrieving data, information, and knowledge for and from humans, and on how to make use of machines for those and other related purposes. Therefore, hardware and software are usually seen more as support and means but, as technology exponentially evolves, there are already many machine learning algorithms, artificial intelligence, and other resources where it is hard for a human mind to fully comprehend the rationale behind its outcomes, results, predictions, processing, or decisions taken, even though they might be shown to be precise and of high quality. There are theoretical e technical efforts to address it, such as the concept of Explainable AI, but it is conceivable that knowledge from machines may not be, in the present or in the future, both efficient and adequately translatable to traditional human-comprehensive knowledge. That knowledge might one day be only usable by other machines in a yet unknown approach of knowledge sharing between them, in a specific way designed for them and perhaps, in the future, also by them: a machine perspective of knowledge and knowledge management. In addition, machine knowledge may not be available only in the explicit form but also in a manner somehow analog to human tacit knowledge, as for instance, a given AI may acquire a rationale that is beyond what its stored bytes can express. That might be also evidence of a context in which perhaps it may be only able to be socialized between machines, in a tacit to tacit “transfer”, not with nor for humans. Furthermore, keeping machine knowledge secure might be far more complex than mere data storage security and policy, as a simple copy of those data may be insufficient for representing and recovering a previously developed machine knowledge, implying that traditional information management is no longer enough. Much is still needed to advance on the topic of machine knowledge, as an approach to data, information, and knowledge from and for machines is needed, in what could be called machine knowledge management (MKM). But that is not the final step needed, as from these machine knowledge and knowledge management concepts emerge the need for a unified theory with human counterparts, that addresses the complex aspects of coexistence and interactions of both clusters of knowledge, with implications for Human-Autonomy Teaming (HAT), and how both can work together in the present and future challenges. Therefore, the aim of this research is to advance toward the proposal of a theoretical model for machine knowledge and knowledge management, on how that can be integrated with the analog human versions in a unified human-machine model, and what might play the mediator role. Subsidiarily, it also discusses the need for a standardized and expanded concept of information and knowledge consistent with that model. Finally, topics are proposed for future research agenda. To achieve these research goals, the main methodologies adopted were the literature review and the grounded theory.