AI-Assisted Generation of Semantic Metadata for Personal Text and Document Files

Problem Statement

"Personal digital text document, such as PDFs, Word-Files, email, and powerpoints, often lack rich contextual metadata. Existing metadata (e.g., file name, creation date) provides only technical details and fails to capture the document’s content, purpose, or personal significance. For digital legacy management, enriched metadata could help preserve meaning, relationships, and life context. Conduct an exploratory user study to identify which types of metadata users find most valuable for organizing and understanding their personal documents. Analyze responses to define metadata categories and prioritize those most relevant to users. Develop a prototype using natural language processing and generative AI models to automatically generate the identified metadata for personal text documents."