My first encounter with AI-generated student work came in a large English 220 course, primarily populated by non-English majors. When the first major papers came in, I noticed the shifts: improved grammar and syntax, the sudden appearance of academic jargon without proper context, less engagement with direct quotes, and generally bland analysis. Running my prompts through AI platforms confirmed my suspicions. My initial reaction was a mix of unpreparedness and disempowerment. However, my primary concern was preserving the rapport and trust I had built with my students.
This experience led to two revelations. First, I realized my role was to be a "doctor, not a detective." My ultimate goal is to help students become better writers, not to catch and punish cheating. The traditional reasons students might plagiarize, pressure, feeling overwhelmed, or insecurity about their abilities, still hold true, even with new technology. This meant accepting what I couldn't control and openly acknowledging my own limitations regarding AI with my students.
Second, I understood the need to "figure it out with my students, not for them." Instead of accusations, I initiated a general discussion about AI, sharing my observations and acknowledging its growing relevance in their lives. I put myself as a co-learner, exploring how this technology could be used effectively. For subsequent assignments, students were required to log their AI usage, and crucially, no one was penalized for initial, undisclosed use.
This collaborative stance allowed me to model responsible AI engagement. I experimented with various platforms myself, demystifying their capabilities and limitations. I demonstrated how AI could be a useful tool for specific parts of the writing process—like generating summaries, brainstorming counterpoints, or creating "shitty first drafts" to overcome writer's block. It could even help condense overly long sentences or provide layman's definitions for complex terms.
However, I also highlighted AI's significant weaknesses. It struggles with the more creative and nuanced aspects of writing, such as deep source integration, synthesizing complex ideas, crafting effective transitions, or developing insightful conclusions. AI often fails to grasp the "sum of parts for the whole," producing generic output that lacks a unique voice. I emphasized that outsourcing the entire writing process to AI could lead to a loss of individual voice and potentially raise questions about their value in professional settings.
To counter this, I doubled down on low-stakes, in-class writing activities. This helped me become familiar with each student's voice, making it easier to spot AI-generated content. More importantly, these exercises provided students with opportunities to practice core writing skills, building their confidence and reducing the temptation to rely inappropriately on AI. For homework, I shifted focus from simple summaries to tasks requiring direct engagement with sources, like identifying and unpacking important quotes—a task AI struggles with.
The outcome was largely positive. Students appreciated the candor, and many who hadn't used AI before became curious about appropriate applications. While some improper use persisted, my established policy allowed for constructive feedback rather than accusatory confrontations. Ultimately, navigating AI in the classroom has reinforced a fundamental truth: the best teaching practices—building rapport, fostering open discussion, and learning alongside students—remain the most effective strategies for preparing them for an evolving world.
This blog is summarized from Hear from Peers Webinar Fall 2024 series