On feedback from bots: Intelligence tests and teaching writing

Abstract

One of the much-debated uses for AI, especially among writing instructors, is the potential for AI to take over the commenting and grading functions of teaching. In this paper, we describe the creation and use of AI for writing feedback in two separate but interconnected approaches: the use of the “Perfect Tutor” exercise in the classroom to teach students to conceptualize the components and priorities we bring to the writing process, and how students might struggle to make use of the same AI for feedback in a less actively guided context, or when the emphasis is not on the metacognition surrounding writing. During our examination of making bots and evaluating their feedback, we explore the limits of current AI. While emphasizing the importance of understanding the limitations, we also identify productive uses of these AI feedback bots in the college writing classroom to develop student critical thinking and writing.

Early versions of parts of this brief article appeared on Inside Higher Ed and Medium (Taylor, 2024; Marino, 2024a; Marino, 2024b).