Skip to main navigation Skip to search Skip to main content

Artificial Intelligence Bias on English Language Learners in Automatic Scoring

  • Shuchen Guo
  • , Yun Wang
  • , Jichao Yu
  • , Xuansheng Wu
  • , Bilgehan Ayik
  • , Field M. Watts
  • , Ehsan Latif
  • , Ninghao Liu
  • , Lei Liu
  • , Xiaoming Zhai

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

Abstract

This study investigated potential scoring biases and disparities toward English Language Learners (ELLs) when using automatic scoring systems for middle school students’ written responses to science assessments. We specifically focus on examining how unbalanced training data with ELLs contributes to scoring bias and disparities. We fine-tuned BERT with four datasets: responses from (1) ELLs, (2) non-ELLs, (3) a mixed dataset reflecting the real-world proportion of ELLs and non-ELLs (unbalanced), and (4) a balanced mixed dataset with equal representation of both groups. The study analyzed 21 assessment items: 10 items with about 30,000 ELL responses, five items with about 1,000 ELL responses, and six items with about 200 ELL responses. Scoring accuracy (Acc) was calculated and compared to identify bias using Friedman tests. We measured the Mean Score Gaps (MSGs) between ELLs and non-ELLs and then calculated the differences in MSGs generated through both the human and AI models to identify the scoring disparities. We found that no AI bias and distorted disparities between ELLs and non-ELLs were found when the training dataset was large enough (ELL≈30,000 and ELL≈1,000), but concerns could exist if the sample size is limited (ELL≈200).

Original languageEnglish
Title of host publicationArtificial Intelligence in Education - 26th International Conference, AIED 2025, Proceedings
EditorsAlexandra I. Cristea, Erin Walker, Yu Lu, Olga C. Santos, Seiji Isotani
PublisherSpringer Science and Business Media Deutschland GmbH
Pages268-275
Number of pages8
ISBN (Print)9783031984617
DOIs
Publication statusPublished - Jul 2025
Externally publishedYes
Event26th International Conference on Artificial Intelligence in Education, AIED 2025 - Palermo, Italy
Duration: 22 Jul 202526 Jul 2025

Publication series

NameLecture Notes in Computer Science
Volume15881 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference26th International Conference on Artificial Intelligence in Education, AIED 2025
Country/TerritoryItaly
CityPalermo
Period22/07/2526/07/25

Keywords

  • AI bias
  • AI disparities
  • Artificial Intelligence
  • Automatic scoring
  • English language learners
  • Science assessment

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Artificial Intelligence Bias on English Language Learners in Automatic Scoring'. Together they form a unique fingerprint.

Cite this