Can AI Level the Playing Field? How AI-Assisted Assessment Impacts Gender Bias in Student Evaluations of Marketing Instructors

Abstract

Marketing instructors increasingly are using artificial intelligence (AI) to improve efficiency in course planning and assessment. However, scholarship has yet to show how such use impacts student evaluations, which are often skewed by gender bias. Toward this aim, we conducted a quasi-experiment in which college-student participants viewed student work along with instructor feedback under four experimental conditions. In one condition, participants were told the grade and feedback were written by the instructor and, in another, that the feedback was generated by an AI tool trained by the instructor. Instructor gender was also manipulated. Participants evaluated the assessment and the instructor using measures of agency and competence. We found that perceptions of agency and competence were higher for male instructors than for female instructors who did not use an AI tool. However, when instructors used AI, marked reductions in gender differences occurred. These findings suggest that AI complicates perceptions of agency and competence enough to potentially level the playing field for female instructors. However, perceptions of instructors were universally more negative when participants were told the assessment was created with AI, regardless of instructor gender. Therefore, attempts to use AI as an assessment tool should be made with caution.