Aisyah, Nurul (57217902481); Al Kautsar, Muhammad Dehan (58882220200); Hidayat, Arif (7006069131); Chowdhury, Raqib (35175634200); Koto, Fajri (56572738100)
Despite rapid progress in vision–language and large language models (VLMs and LLMs), their effectiveness for AI-driven educational assessment in real-world, underrepresented classrooms remains largely unexplored. We evaluate state-of-the-art VLMs and LLMs on over 14K handwritten answers from grade-4 classrooms in Indonesia, covering Mathematics and English aligned with the local national curriculum. Unlike prior work on clean digital text, our dataset features naturally curly, diverse handwriting from real classrooms, posing realistic visual and linguistic challenges. Assessment tasks include grading and generating personalized Indonesian feedback guided by rubric-based evaluation. Results show that the VLM struggles with handwriting recognition, causing error propagation in LLM grading, yet LLM feedback remains pedagogically useful despite imperfect visual inputs, revealing limits in personalization and contextual relevance. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2027.
Quantic School of Business and Technology, Washington, DC, United States; Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates; Indonesia University of Education, Bandung, Indonesia; Monash University, Melbourne, Australia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock