AI Feedback as Revision Support

Students in online writing courses often make important revision decisions on their own time, between instructor interactions and before formal feedback may be available. In that space, a useful question emerges: can AI-supported feedback help students think more intentionally about their writing without taking the writing out of their hands?

Most AI-based writing feedback is often imagined as correction-driven: a quick scan, a list of fixes, and maybe a cleaner sentence at the end. And there are several widespread “grammar” tools that function in this way as well – click to fix, no thought required. That kind of support can be useful in limited ways, but it does not necessarily help students understand their writing or make better revision decisions. In writing instruction, students need feedback that is developmental rather than merely corrective. That means feedback that recognizes what is working, identifies areas for growth, explains why those areas matter, and helps students decide what to do next. For AI feedback to support learning, it has to be designed for that kind of thinking, not just for speed or surface-level correction.

A recent pilot study in ENG121 explored that question through a custom AI-based feedback tool designed for first-year writing students in online, asynchronous courses. The tool was optional and was aligned with course outcomes, assignment expectations, and rubrics. Just as important, it was intentionally constrained: it did not rewrite student sentences or produce revised passages. Instead, it offered developmental feedback focused on rhetorical purpose, organization, idea development, and possible next steps for revision.

Student responses suggest that AI feedback can be useful when it helps writers understand what to improve and how to think about revision. Several students described the feedback as actionable and developmental, noting that it helped them identify areas for improvement, review drafts before submitting, and make decisions about next steps. Students also used the tool differently, some earlier in the drafting process and others closer to submission, suggesting that optional tools may support different writing habits and levels of confidence.

For faculty, the takeaway is not simply that AI can give feedback. It is that feedback tools need instructional design. That applies well beyond writing. Any course that asks students to practice a skill, interpret expectations, revise their work, solve problems, apply feedback, or make judgments could benefit from this same design logic. The starting point is identifying where students need better support while they are still learning.

A faculty member considering a similar tool might begin by identifying a specific learning moment: reviewing a draft, checking alignment with assignment criteria, preparing for a lab report, practicing case analysis, interpreting feedback on a project, or testing understanding before submission. From there, the design questions become clearer: What should the tool help students notice? What should students remain responsible for? What should the tool avoid doing? How will students understand its role alongside instructor feedback, rubrics, examples, course materials, and their own judgment?

In that sense, the broader implication is not that every course needs an AI tool. It is that faculty can shape AI-supported learning by defining the task, setting the boundaries, and keeping student thinking at the center. The most useful tools may be the ones that do less than students expect but do that smaller job in a way that helps them learn.

AI feedback is most promising when it supports the learning process rather than short-circuiting it. Designed carefully, it can create one more opportunity for students to pause, interpret feedback, and revise with purpose.

Read the full article.

Reference
Pritts, N. (2026). Student perspectives on AI-based feedback in first-year writing: Perceptions and instructional implications. International Journal of Innovative Teaching and Learning in Higher Education, 7(1). https://doi.org/10.4018/IJITLHE.413093

Dr. Nathan Pritts

Bio: Nathan Pritts has worked in higher education for more than 20 years as an educator, program leader, scholar, and institutional strategist. He currently serves as Principal AI Strategist and Professor at the University of Arizona Global Campus, where he leads cross-functional work spanning academic innovation, operational workflows, student success, and organizational change. Grounded in large-scale online learning environments, his work focuses on helping institutions move from experimentation to sustainable, human-centered practice. Prior to his AI strategy role, Pritts served as Program Chair of First-Year Writing, leading one of the university’s largest academic units and overseeing curriculum, assessment, faculty development, and instructional quality across high-enrollment general education courses. He is the editor of Empowering Educational Development and Faculty Growth With AI and the author or co-author of fourteen books, including Film: From Watching to Seeing, Essentials of Academic Writing, and the forthcoming AI as a Creative Partner: An Introduction from Routledge.