Identifying Fairness Issues in Automatically Generated Testing Content

Kevin Stowe, Benny Longwill, Alyssa Francis, Tatsuya Aoyama, Debanjan Ghosh, Swapna Somasundaran


Abstract
Natural language generation tools are powerful and effective for generating content. However, language models are known to display bias and fairness issues, making them impractical to deploy for many use cases. We here focus on how fairness issues impact automatically generated test content, which can have stringent requirements to ensure the test measures only what it was intended to measure. Specifically, we review test content generated for a large-scale standardized English proficiency test with the goal of identifying content that only pertains to a certain subset of the test population as well as content that has the potential to be upsetting or distracting to some test takers. Issues like these could inadvertently impact a test taker’s score and thus should be avoided. This kind of content does not reflect the more commonly-acknowledged biases, making it challenging even for modern models that contain safeguards. We build a dataset of 601 generated texts annotated for fairness and explore a variety of methods for classification: fine-tuning, topic-based classification, and prompting, including few-shot and self-correcting prompts. We find that combining prompt self-correction and few-shot learning performs best, yielding an F1 score of 0.79 on our held-out test set, while much smaller BERT- and topic-based models have competitive performance on out-of-domain data.
Anthology ID:
2024.bea-1.20
Original:
2024.bea-1.20v1
Version 2:
2024.bea-1.20v2
Volume:
Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024)
Month:
June
Year:
2024
Address:
Mexico City, Mexico
Editors:
Ekaterina Kochmar, Marie Bexte, Jill Burstein, Andrea Horbach, Ronja Laarmann-Quante, Anaïs Tack, Victoria Yaneva, Zheng Yuan
Venue:
BEA
SIG:
SIGEDU
Publisher:
Association for Computational Linguistics
Note:
Pages:
232–250
Language:
URL:
https://aclanthology.org/2024.bea-1.20
DOI:
Bibkey:
Cite (ACL):
Kevin Stowe, Benny Longwill, Alyssa Francis, Tatsuya Aoyama, Debanjan Ghosh, and Swapna Somasundaran. 2024. Identifying Fairness Issues in Automatically Generated Testing Content. In Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024), pages 232–250, Mexico City, Mexico. Association for Computational Linguistics.
Cite (Informal):
Identifying Fairness Issues in Automatically Generated Testing Content (Stowe et al., BEA 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.bea-1.20.pdf