Student Decision-Making in AI-Supported Writing
Noreen Fitzgerald-Makar, Queensborough Community College
Bio
Noreen Fitzgerald-Makar has been an adjunct lecturer through the College Now program since 2011; she primarily teaches EN101 Introduction to College Composition. In NYCPS she is the ELA and journalism teacher at Thomas A. Edison CTE High School where she has worked since 2007. Since 2011, she has advised the student news publication The Edison Light. She works with the Journalism for All and Press Pass NYC youth journalism initiatives in New York City.
Course Setting
I teach EN101 on two campuses through the College Now program. The primary class is taught on Saturdays at Queensborough Community College Campus from 9am -1pm; the students that form this class come from a variety of Queens High Schools, including Francis Lewis, Bayside, Cardozo, Edison, Bowne, and Van Buren. Classes are composed of 11th and 12 graders of a wide range of cultural backgrounds.
Problem of Practice
My goal was to explore how AI could be integrated into the editing process in a way that still supported student agency and critical thinking. I wanted to intentionally scaffold students’ use of AI so they could make thoughtful, informed, and ethical decisions about revision suggestions rather than relying on AI in ways that diminished their individual voice and writing style. Focus question: How might I use guided AI-supported editing practices in my EN101 dual-enrollment classroom to help students critically evaluate AI suggestions and engage in deeper revision rather than passively accepting edits?Strategy
The strategy focused on guiding students’ intentional use of AI during the editing process in order to preserve student voice and strengthen metacognitive awareness during AI-assisted editing and revision. Initial steps included students completing a baseline writing assessment (personal narrative) to establish independent writing style and revision habits, along with a pre-survey measuring students’ prior experiences, attitudes, and comfort levels with AI-supported writing.
Students then participated in guided practice analyzing ineffective and effective AI prompts for revision, particularly prompts related to grammar, organization, clarity, transitions, and voice preservation, before creating their own prompts to revise their Compare/Contrast essays. Following revision, students reflected on why they accepted, rejected, or modified specific AI-generated suggestions and how those decisions impacted their essays.
Students repeated the same process during the next writing cycle for their argumentative essays and tracked their prompts, revision decisions, and reflections in a shared Google Sheet. Analysis of student reflections revealed recurring concerns about AI reducing individual voice, creativity, authenticity, and intentionality in student writing. Students view AI as a highly effective revision assistant for clarity and mechanics, but they are still developing confidence and strategies for evaluating its suggestions.
Documentation
Student Prompts and Responses Cycle 1
This artifact shows samples of student-created prompts meant to help outline and draft their compare/contrast essays on individually proposed topics. Students explained what suggestions they accepted and rejected from AI and reflected on choices made and why.
Student Prompts and Responses Cycle 2
Samples of student-created prompts to help outline and draft their argumentative essays on individually proposed topics. This assignment required source material and several students tried to prompt AI to generate sources–some were initially skeptical to accept the AI-generated suggestions.
Measuring Impact
AI Suggested Revisions
This graph describes the changes in student responses to the prompt “AI suggestions helped me notice areas of my writing that needed revision.” Between cycle one and cycle two the number of students that replied “neutral” decreased and the number that responded “strongly agreed” was reduced to zero.
Process and Intention
This graph shows student responses to the prompt “The process helped me think more intentionally about why I was making changes to my essay.” The number of students that replied “strongly agree” doubled from my initial implementation, while the number that disagreed or strongly disagreed was reduced to zero.
Analysis
The process was meaningful but required more class time than expected, as students often did not complete AI-prompting work outside of class. Despite this, both cycles were successful in that students did not submit drafts that “looked like they had been run through AI” or sounded overly automated.
Overall, student reflections show that AI was most effective when used as a revision support tool rather than an editor. Students consistently described using AI for grammar, transitions, structure, and clarity while still maintaining their writing voice. As one student noted, AI helped with “grammar, spelling errors, feedback on how to make it stronger,” while others explained that suggestions improved “transitions… without changing my writing style.” Students frequently rejected suggestions that altered tone or removed voice, explaining that they wanted to keep writing that “still made it sound like me.”
Students reported that AI strengthened their argumentative writing; one reflection noted that AI “told me other ways to write my counter argument” and helped “make my claim even stronger.” Others emphasized improvements in structure, explaining that AI provided “a clear structure for organizing my argumentative essay” and helped reduce repetition and run-on sentences.
Across both cycles, students believed they were already relatively skilled with AI, but the data suggests a meaningful shift in how intentionally they engaged with it. Post-survey responses indicate that students increasingly viewed AI as a tool for support rather than an automatic editor. This shift is reflected in their willingness to evaluate suggestions and selectively apply feedback. While there is still some variability in student responses, overall there is a clear pattern of increased metacognitive awareness and decision making.
Recommendation
Transparent AI use as a supportive tool was helpful in developing student metacognition and teachers should have a candid discussion with students at the start of the class about the use of AI as a supportive tool and how to best do this ethically. From this discussion teachers should model prompt writing and show students that it is a process; no one prompt is a “magic bullet,” it is a multi-step process that involves critical thought and reflection. Making it clear from the start that students should not allow AI to make edits, but offer suggestions, was very helpful in order to help students make their own choices with their work and to recognize that their individual voices are more impactful than they give themselves credit for.
Resources
- Student Survey: AI-Supported Revision in EN101
- AI Prompt Tracker Comp/Cont and Argument Student Responses
- Rubric: Intentional AI-Supported Revision (EN101)
- AI Ethics and Hallucinations
- Evaluating Information: Tip: Generative AI