LLMs Do Not Grade Essays Like Humans

2026-03-24

AI Summary

This study addresses the unclear alignment between large language models (LLMs) and human raters in automated essay scoring. It presents the first systematic evaluation of off-the-shelf GPT and Llama family models in a zero-shot setting, combining correlation analysis with error pattern identification. The findings reveal that these models rely on different signals than human graders—tending to overrate short essays and underrate longer ones containing minor grammatical errors. Although model scores show only weak agreement with human ratings, they exhibit strong internal consistency with the qualitative feedback the models generate, suggesting potential utility as assistive grading tools. This work provides critical empirical evidence for understanding the scoring mechanisms of LLMs and their implications for educational applications.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large language models have recently been proposed as tools for automated essay scoring, but their agreement with human grading remains unclear. In this work, we evaluate how LLM-generated scores compare with human grades and analyze the grading behavior of several models from the GPT and Llama families in an out-of-the-box setting, without task-specific training. Our results show that agreement between LLM and human scores remains relatively weak and varies with essay characteristics. In particular, compared to human raters, LLMs tend to assign higher scores to short or underdeveloped essays, while assigning lower scores to longer essays that contain minor grammatical or spelling errors. We also find that the scores generated by LLMs are generally consistent with the feedback they generate: essays receiving more praise tend to receive higher scores, while essays receiving more criticism tend to receive lower scores. These results suggest that LLM-generated scores and feedback follow coherent patterns but rely on signals that differ from those used by human raters, resulting in limited alignment with human grading practices. Nevertheless, our work shows that LLMs produce feedback that is consistent with their grading and that they can be reliably used in supporting essay scoring.

Problem

• automated essay scoring
• large language models
• human grading alignment
• essay evaluation
• LLM-human disagreement

Innovation

• automated essay scoring
• large language models
• human-LLM alignment
• grading behavior
• feedback consistency