Logo image
Evaluating the role of AI in writing assessment development through the lens of the validity argument
Journal article   Peer reviewed

Evaluating the role of AI in writing assessment development through the lens of the validity argument

Catherine Welch, Stephen Dunbar and Yen Vo
Assessing writing, Vol.70, 101107
10/2026
DOI: 10.1016/j.asw.2026.101107

View Online

Abstract

Artificial intelligence (AI) holds tremendous potential for the field of education, from improving accessibility to personalizing learning to measuring outcomes. Yet, it is imperative that the integration of AI into writing assessments is based on research and validity evidence to support appropriate purposes and use. This paper will address validity evidence that is necessary to evaluate AI-driven models that contribute to development and scoring activities of writing assessments. This abstract poses questions that the paper will address as types of validity evidence that contribute to Kane’s (2006) interpretive use argument. •The test development process should be validated for any new AI innovation.•The foundational concepts of validity should be paramount in the evaluation of the AI contributions.•AI implementation needs to be well documented and monitored for its own quality control.
Generative AI Test development procedures Validation

Details

Logo image