AI Insight
This study develops a mathematical model to analyze how knowledge spreads within organizations when employees learn both from each other and from AI systems. The researchers found that knowledge dissemination exhibits a threshold effect where it either spreads successfully or decays, and that AI-enabled mechanisms significantly improve knowledge transformation by enhancing both efficiency and contact intensity between learners. Using a five-stage employee classification system and cellular automaton simulations, they demonstrate that optimal control strategies can improve knowledge coverage while maintaining system stability under resource constraints.
Why it matters
The findings provide organizations with theoretical guidance for designing AI-assisted knowledge management systems that balance human interaction with machine learning tools. The identified threshold effects and optimal control strategies offer practical frameworks for companies investing in digital transformation to maximize employee knowledge acquisition and retention.
Understand the Science
by Yanhua Lv, Yadong Wang, Huanhuan Wang, Sida Kang
With the deep integration of digitalization and artificial intelligence technologies into enterprise knowledge management, AI-assisted learning is reshaping the channels through which knowledge is acquired within organizations. Meanwhile, knowledge dissemination is evolving from traditional interpersonal interaction toward human-machine collaborative interaction. In light of this transformation, this study develops a multi-stage 2SEIR–CA dynamic model of knowledge dissemination under an AI-enabled mechanism that incorporates dual-channel “interpersonal–human-machine” interaction. Specifically, employees are classified into five states: employees who have not yet mastered the knowledge, employees with limited knowledge mastery, employees with strong knowledge mastery, knowledge-disseminating employees, and knowledge-exited employees. On this basis, the next-generation matrix method is employed to calculate the basic reproduction number of the system and to analyze the existence and stability of equilibrium points. An optimal control model is then established to derive strategies for promoting knowledge dissemination, and a cellular automaton model is introduced to characterize its spatiotemporal evolution. The results show that the system exhibits a pronounced threshold effect: when key dissemination parameters exceed their critical levels, knowledge can continue to diffuse within the organization; otherwise, it will gradually decay. Second, the AI-enabled mechanism enhances knowledge transformation efficiency and contact intensity, thereby significantly promoting the transition of employees with strong knowledge mastery to knowledge-disseminating employees. Third, the optimal control strategy proposed in this study can optimize transformation pathways across different stages under resource constraints, thereby improving overall dissemination efficiency and maintaining the structural stability of the dissemination system. Fourth, the spatiotemporal simulations demonstrate that, under optimal regulation, knowledge dissemination can form a stable diffusion pattern with high coverage. Under the assumptions and parameter settings of the proposed model, the findings provide theoretical insights into the dynamic mechanisms underlying AI-enabled knowledge dissemination and offer model-based implications for knowledge management practices in organizations.