By LISA VOTODIAN
In a previous University Times article, I wrote about the importance of feedback literacy skills for both instructors and students and how these literacies are closely linked. I also talked about the challenges students and instructors face in effectively giving and receiving feedback and why more feedback isn’t necessarily better.
In this article, I’d like to examine how a multi-source feedback approach — strategically integrating input from instructors, peers and generative AI — can create a comprehensive feedback ecosystem that fosters students’ ability to use feedback constructively, develop awareness of the roles of instructors and learners, and effectively manage emotional responses to feedback.
Alan Lesgold and John Radzilowicz introduced The Conductor’s Paradox to explain that just as “a symphony conductor must deeply understand music’s structures before effectively directing musicians … students must develop core abilities through direct practice before they can orchestrate AI tools effectively.”
Foundational cognitive development and specific competencies are at risk when AI is not introduced sequentially and at appropriate times. The comprehensive feedback ecosystem model builds on this concept by starting the feedback sequence with AI precisely because it is lower-stakes emotionally and allows students to build skill and confidence before facing peer or instructor judgment.
Why feedback often fails students: The feedback-literacy gap
Students encounter complex challenges in developing feedback literacy. They may have difficulty interpreting academic jargon or were never taught how to effectively implement feedback for improvement. Assumptions, expectations, inherent bias and perceptions of power can hinder a student’s ability to confidently assume an active role in the feedback process. Social and cultural background, year of study, and course format also influence how students learn, perceive feedback and interact with faculty.
More feedback isn’t necessarily better. Lesgold and Radzilowicz’s broader argument about foundational cognitive development provides the rationale for why emotional scaffolding in feedback matters: If students haven’t built the metacognitive capacity to interpret, react to and apply feedback, more feedback (of any kind) won’t help.
Benefits of a multi-source feedback approach
By harnessing multiple feedback sources in a comprehensive, developmentally based ecosystem, you leverage the best that each type of feedback can provide, minimize some of the key drawbacks of each form of feedback, and acclimate students to the concept of feedback as a dialogue that guides students and their work toward a specific goal.
A scaffolded approach helps students cultivate emotional maturity and resilience to receive critical feedback openly, reflect on their performance, and engage productively with evaluative input. Each stage of the feedback ecosystem introduces a complexity that students are ready for letting them reinforce mastery while gaining confidence and continuing to build on skills that will better prepare them for more effective instructor interactions.
AI-generated feedback. A scaffolded approach — beginning with AI generated feedback tools that give students immediate, frequent feedback during formative stages — helps them learn to manage emotions triggered by feedback, begin to make judgments about their work, and use feedback to improve. AI feedback tools can be designed so they don’t replace student writing or thinking — they respond to it. Students still do the cognitive work; the AI reacts.
Peer feedback. Once students begin to cultivate emotional maturity and resilience in receiving critical feedback, they can move on to a more emotionally complex peer feedback source. Peers introduce social connectivity through relatedness and perspective sharing. Students see it as an effective social learning construct in increasing engagement and motivation and providing exposure to multiple perspectives and varying levels of standards and exemplars.
Instructor feedback. Instructors play a key role by providing expert validation, feedback on higher-order thinking, and insights on critical thinking, content accuracy, and creativity. They can strategically monitor and intervene during the AI and peer feedback stages to provide support when misconceptions or emotional needs arise.
How to implement the approach
Creating a comprehensive feedback ecosystem will take time, and the model only works if the roles are kept distinct, and students understand why.
1. Develop a course-customized framework with the help of AI. Claude or ChatGPT can help create a feedback framework that includes AI, peer, instructor, and self-assessment by uploading assignment criteria, rubrics, and a timeline. Then ask it to suggest or design an AI-generated feedback tool for students to engage with. Refining prompts by adding context, offering examples, and specifying output constraints allows you to customize the feedback tool. For instance, you can add a Socratic element that prompts students to explore their own understanding by challenging assumptions, justifying choices, and explaining reasoning through iterative dialogue, creating deeper learning, and critical thinking.
2. Establish a conceptual foundation for feedback literacy. Active discussion, collaborative exploration and experiential activities reframe feedback as a learning tool rather than as negative judgment, creating psychological safety and shared understanding before any feedback is given or received. For example:
Use a pre-course survey, self-reflection exercise, or small discussion groups and ask students to share their experiences with receiving feedback in the past.
Discuss the results, recognizing emotional reactions and challenges in interpreting and applying feedback and introducing the concept of feedback literacy — the skills and mindsets to seek out, understand and use feedback effectively.
Be transparent and provide a copy of the feedback framework that explains each source of feedback students should expect to receive and why each is important.
3. Introduce AI tools deliberately. Faculty guidance significantly shapes students’ perspectives and uses of AI tools. Providing clear direction in using AI for feedback purposes, helping students critically assess AI-generated suggestions, and providing feedback dialogue opportunities are key to mitigating AI’s limitations, establishing effective feedback cycles, and cultivating feedback skills.
Explain the purpose and role AI plays, when it’s permitted or encouraged and why.
Distinguish between appropriate tool use and cheating and what happens with inappropriate use.
Whether/how students must acknowledge GenAI use.
Select the AI tools you’ll use: For example, Grammarly provides real-time feedback on grammar, spelling, punctuation, and tone. Users can also receive suggestions for improving clarity and structure and even explore potential counterarguments for academic writing. QuillBot features a free essay and paper checker that scans for grammar, spelling, punctuation, and fluency errors. It helps users learn from their mistakes and offers tools for paraphrasing and summarizing content.
Provide alternative feedback sources for students who do not want to use AI in any part of their coursework.
Anonymize personal, student, or University data before uploading it to an AI tool.
Never use AI to assign grades.
4. Coach and mentor students in the peer feedback process. Designing peer feedback activities and discussions effectively involves a thoughtful approach to ensure that they are constructive, balanced, and conducive to learning.
Provide feedback criteria and moderate peer feedback.
Offer prompts to assist students in giving feedback.
Conduct a mini workshop on constructive feedback techniques.
Peerceptiv’s peer review feature builds students’ writing, critical thinking, and content knowledge simultaneously.
5. Instructor feedback. Traditional models put the instructor at the end of the process — as a final judge. Redistributing earlier feedback to AI and peers allows instructors to intervene strategically at multiple points rather than exhaustively at one and provide deeper insights on critical thinking, content accuracy, and creativity.
Do you have an assignment in mind that this model would work well with? Schedule a consultation with the Teaching Center to collaborate with one of our teaching consultants on building a comprehensive feedback ecosystem. Don’t forget to share your story with us.
AI tools were used to sharpen the structure and language of this article.
Lisa Votodian is the survey assessment manager of the Office of Measurement and Evaluation of Teaching at the University Center for Teaching and Learning
REFERENCES AND RESOURCES
Ardill, N. (2025). Peer feedback in higher education: student perceptions of peer review and strategies for learning enhancement. European Journal of Higher Education, 15(4), 696–721. https://doi.org/10.1080/21568235.2025.2457466
Little, T., Dawson, P., Boud, D., & Tai, J. (2025). What does it take to provide effective peer feedback? Assessment & Evaluation in Higher Education, 50(5), 775–786. https://doi.org/10.1080/02602938.2025.2475059
Pitt, E., & Norton, L. (2017). ‘Now that’s the feedback I want!’ Students’ reactions to feedback on graded work and what they do with it. Assessment & Evaluation in Higher Education, 42(4), 499–516. https://doi.org/10.1080/02602938.2016.1142500
Xu, Q., Liu, Y., & Li, X. (2025). Unlocking student potential: How AI-driven personalized feedback shapes goal achievement, self-efficacy, and learning engagement through a self-determination lens. Learning and Motivation, 91, 102138. https://doi.org/10.1016/J.LMOT.2025.102138
Zhan, Y., Boud, D., Dawson, P., & Yan, Z. (2025). Generative artificial intelligence as an enabler of student feedback engagement: a framework. Higher Education Research & Development, 44(5), 1289–1304. https://doi.org/10.1080/07294360.2025.2476513
AI-Enhanced Feedback Loops – By Caryn Sever – a Canvas Commons module
Visit the Anthropic Academy to learn more about Claude and creating artifacts like AI generated feedback tools.