Evaluating Handwritten Student Work in Latvian Using LLMs’ General Knowledge

Artificial Intelligence (AI) has now entered all areas of education and is creating many new opportunities for research, new approaches to learning and innovation, but it is also posing challenges for the educational environment. There are: breaches of academic integrity, under-learning, problems with assessment, hallucinations caused by AI. During this study the prototype system utilizing a Large Language Model (LLM) was developed and tested to analyse the results and find out if it can be used to assess students’ handwritten work according to set criteria. The study has shown that it is possible to use LLM-based system to assess students’ handwritten work to ensure that students’ hand written responses are a real reflection of their learning, but further work is needed to ensure that assessments of the model and educator are as equal as possible. By exploring LLM’s role in automating exam evaluation and grading, this study contributes to the advancement of AI-powered personalized learning tools. The findings have implications for reducing the burden on educators, improving assessment efficiency, and fostering adaptive learning environments.