Tool for On-Demand Virtual Courses Generation using Content Curation and Artificial Intelligence

Del Carpio A.F.; Anccasi C.M.; Alvarado A.A.; Arosquipa L.M. · Proceedings of the IEEE International Conference on Software Engineering and Service Sciences, ICSESS (2025) · DOI: 10.1109/icsess67729.2025.11380731

Resumen

The continuous production of tutorial videos on the internet about several topics provides a rich source of educational material that should be leveraged effectively. This generates a need for tools that enable easy access to educational content for learners interested in self-directed study. This paper introduces an innovative tool for dynamically creating an on-demand course. It addresses challenges such as information overload and content organization by using YouTube's extensive video library combined with the advanced processing and generation capabilities of Large Language Models (LLMs) and content curation techniques. The functions of the tool enable, firstly search and selection of YouTube videos based on user prompts. Then, the Gemini-2.5-flash model analyzes the selected videos and metadata to generate a coherent course structure and complementary material. Finally, a crucial LLM-assisted content curation phase refines and structures the content, seeking to align with the subjects of each section. This approach provides a personalized learning path, optimizing accessibility by obtaining targeted material from YouTube videos for effective learning. Initial evaluations demonstrated partial results of the tool to produce structured course content that meets learners' needs. Although with initial results, this tool can establish a promising direction for democratizing access to personalized learning.

Instituciones: Universidad La Salle