Assessment of Short Essay Questions by LLMs:
Challenges and Potential Educational Applications
DOI:
https://doi.org/10.21814/lm.18.2.509Keywords:
Automatic short-answer grading, ASAG, evaluation of student answers, large language models, LLMAbstract
The relationship between emerging technologies and human language has been widely investigated, especially after the popularization of large language models (LLMs) and their ability to interpret and generate texts. In educational settings, there is intense debate about the use of these technologies in classrooms by students, teachers, and institutions. Considering recent advances in the use of Artificial Intelligence (AI) in computational linguistics, this research is motivated by the need to integrate innovative methodologies into textual analysis, specifically in the assessment of short-answer questions (ASAG, Automatic Short Answer Grading) answered by undergraduate students in a text interpretation activity. This study compares human evaluation, marked by subjectivity, with the capabilities of LLMs, whose operation is based on probabilistic calculations. Data analysis revealed a very similar grading pattern between the two human evaluators, whereas the LLMs showed greater variation, both across models and in comparison with human evaluators. The highest positive correlation between an LLM and human evaluators was observed for Gemini 3, in one of the prompt variations tested in this study. These findings indicate that further research is still needed to refine tools for the ASAG task in Portuguese. Nevertheless, the observed performance of models such as Gemini suggests potential educational contributions, particularly in supporting teachers' diagnosis and feedback practices in short-answer assessment.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Roberlei Alves Bertucci, Osmar de Oliveira Braz Junior, Renato Fileto

This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).








