Automating Bias in Writing Evaluation
Maria Goldshtein, Amin G. Alhashim, Rod D. Roscoe · 2024
Writing assessment is a ubiquitous component of education, job seeking and promotion, everyday communication, and more. To serve these crucial activities, automated writing evaluation (AWE) technologies have been developed to facilitate and streamline assessment and feedback by providing faster, more consistent, and ostensibly less biased judgments than human evaluators. This chapter synthesizes insights from computer science, sociolinguistics, social psychology, and ethics to (a) highlight the potential for bias inherent to the building and implementation of AWE along with barriers to resolving those issues, and (b) discuss research on bias mitigation strategies. This chapter is intended to propel future research toward more equitable development and use of AWE.