In the fall of 2023, I noticed something unusual in my students’ essays. Their work seemed to split into two distinct categories. One group submitted papers that were well-polished, particularly in grammar, and several students admitted during conferences that they heavily relied on AI tools, sometimes even for entire drafts. Surprisingly, many did not think they had done anything wrong. The other group submitted rough, underdeveloped essays that were challenging to follow. I often thought these students could have benefited from using AI tools before meeting with me. This contrast led me to wonder: if students were taught to use AI more critically, could it enhance their writing rather than replace it? Instead of banning or ignoring AI, I decided to teach students to use it as an assistant during revisions.<\/p>\n
Duri Long and Brian Magerko (2020) describe AI literacy as a set of skills that allow people to evaluate AI critically, communicate with it effectively, and use it as a tool in various settings. In my classroom, this definition explained what I observed: students used AI frequently but not always thoughtfully. They struggled less with access to AI and more with using it intentionally. Most students reported using AI tools, but their confidence in using them effectively was lower. This gap indicated that familiarity with AI did not guarantee critical or ethical use. It became evident that students needed guided practice, not just rules. I designed a revision project that framed AI as an assistant, requiring students to decide when to accept, revise, or reject its suggestions.<\/p>\n
To address these issues, I created an assignment treating AI as a revision assistant, not a substitute. The aim was for students to think critically about AI feedback, consider its ethical use, and make intentional choices about which suggestions to accept or reject. Students revisited a Character Analysis Research Paper, running it through the Hemingway Editor, which focuses on sentence clarity and readability. They reviewed three to eight suggested changes, choosing to accept, revise, or reject them. The tool also allowed students to adjust the reading level of their drafts, prompting discussions about oversimplifying ideas and maintaining a student’s voice. Unlike generative tools, Hemingway offers options rather than producing content. After revising, students compared their drafts and wrote reflections on how the AI tool influenced their writing.<\/p>\n
Before starting the assignment, students showed varied preparation in AI use. Many could distinguish between generative and assistive tools, while others were less familiar with AI writing tools. Those with experience mentioned platforms like Grammarly or ChatGPT, indicating that while AI wasn’t new, understanding when and how to use it effectively varied. Students expressed concerns about AI generating incorrect information, diminishing their voice, or crossing ethical lines close to cheating. These fears mirrored instructors’ worries that AI might shortcut learning. The assignment allowed students to confront these fears through guided use and reflection.<\/p>\n
Despite concerns, many students found assistive tools helpful during revisions. They reported improvements in grammar, clarity, and sentence length, and noticed patterns in their writing they had overlooked. Some resisted the tool’s simplistic suggestions, yet many made meaningful revisions, selecting suggestions that aligned with their goals. This selective engagement indicated growing critical judgment. While the tool didn’t drastically change performance, it raised a question for me: how might outcomes shift if AI were introduced earlier in the writing process? These mixed reactions reinforced the importance of when and how students use AI.<\/p>\n
Based on insights from the first semester, I revised the assignment for Spring 2025. Although students found the AI assistant beneficial, they felt overwhelmed by the feedback volume. I continued using Hemingway Editor as it required students to evaluate and select suggestions. The second version of the assignment incorporated structured, in-class time for using Hemingway Editor. Students began with demonstrations and independent practice, using AI before submitting final drafts. I added a survey question to see if students’ views on AI in academics changed after the assignment.<\/p>\n
Comparing the two project versions highlighted the impact of instructional framing on student behavior. When AI use wasn’t graded, students experimented more freely but casually. Once integrated into the graded writing process, students became more selective, consulting and questioning the tool rather than following it blindly. This shift underscored the importance of AI’s role in the classroom. Introducing it earlier made students engage more deliberately with revisions aligned with their purpose and voice. Modeling use also mattered. By demonstrating how to work through the tool, students grew more confident in making their own decisions. In future assignments, I plan to have students try multiple AI tools or different ones at various writing stages.<\/p>\n