Generative AI-Based Reading Literacy Assessment System for Independent Learning in Resource-Constrained Educational Environments

Abiyasa Putra Prasetya¹, Yan Watequlis Syaifudin¹, Rakhmat Arianto¹, Mustika Mentari², Triana Fatmawati¹, Pramana Yoga Saputra¹

1. ¹ Politeknik Negeri Malang
2. ² Okayama University

International Conference on Vocational Education and Electrical Engineering (ICVEE), 2026

Reading literacy assessment is essential to support critical thinking, knowledge acquisition, and independent learning, yet many educational environments continue to face challenges related to low literacy achievement, limited resources, and difficulties in providing timely formative feedback. Recent advances in generative artificial intelligence (AI), particularly large language models (LLMs), offer new opportunities to automate assessment and support personalized learning. This paper presents a generative AI-based reading literacy assessment system that integrates automated question generation, semantic-based essay evaluation, and formative feedback generation using a locally deployed LLM and a Retrieval-Augmented Generation (RAG) framework. The proposed architecture operates independently of cloud services, making it suitable for resource-constrained educational environments while preserving accessibility, privacy, and sustainability. The system was evaluated using Indonesian folklore texts from official literacy learning materials. Experimental results demonstrate its ability to generate contextually relevant assessment items and provide consistent automated evaluation and feedback, highlighting the potential of local AI deployment for literacy assessment and independent learning.

DOI10.1109/icvee71133.2026.11766322Copy