Implementation Research on Constructing Reliable Information Stocks in Digital Platforms: An Empirical Analysis of the Multimodal Persona Interface (MPI) and a 3,000-Article Dataset

Abstract This study investigates an implementation-based framework for constructing reliable, stock-oriented information assets within digital ecosystems characterized by the attention economy. The research addresses the rapid obsolescence of information and the resulting “intellectual blank spaces” caused by educational and regional disparities. Over a 14-month period, 3,000 articles were published on a digital platform (note, Japan), garnering 220,000 views and 4,000 user endorsements. We propose the Multimodal Persona Interface (MPI), a framework that integrates visual symbols, dialogic explanations, diagrams, audio commentary, and multilayered summaries to reduce cognitive load and enhance structural understanding. Analysis of viewership logs identifies four stable demand domains: survival-related information, social verification, learning support, and cultural interpretation. The results suggest that individual knowledge production, when structured via MPI, can function as a resilient public information infrastructure. Keywords: Multimodal Persona Interface (MPI), Information Stock, Attention Economy, Digital Divide, Knowledge Infrastructure, LLM-Assisted Content Engineering. Conflicts of Interest: The author declares no conflicts of interest associated with this research.