At the beginning of this school year, our staff set out to craft a fresh AI policy. We found that our existing plagiarism policies didn’t cover certain situations. For instance, how should students handle an increasingly assertive Grammarly extension that now wants to do way more than fix basic grammar and spelling? Should students accept every “suggestion” that the extension offers? Is it cheating if a student clicks through the options under, “How this may sound to readers” and then selects “confident?” Is it cheating if a student clicks, “Use this version” after Grammarly “polishes” an entire paragraph? At what point does a piece of writing (or artwork, or code) stop being the student’s work?
We split into departments to discuss the issue and craft new policy language. As the ELA team gathered, I expected us to spend most of our time untangling the Grammarly conundrum. But an interesting thing happened: Our conversation turned away from what happens after students have written drafts. Instead, we kept being drawn to an issue that seems to attract much less attention than AI plagiarism: students’ use of AI to start the writing process.
The New Path of Least Resistance
Many of us have noticed that the first thing stumped writers think to do is ask a generative AI model for story ideas. Until a year or so ago, these same students would have stared at a blank screen or notebook for a while, then spoken to a classmate or asked their teacher for advice. The friend might have shared what they were writing. The teacher might have asked the student to look back to a key moment from the mentor text for inspiration, or even nudged them to remember the storyline from a favorite book, video game, or TV show. The student might have looked around the pre-AI version of Google for ideas. Clicking links would have brought them to other pieces of writing, along with relevant images, videos, music. This process would have been a struggle. But it was also, our department agreed, often a productive struggle (Sriram, 2020). Through it, students got the opportunity to practice self-advocacy and perseverance. They also learned to connect current writing challenges to those they (and others) have already faced and overcome.
There’s something special about students finishing narratives *they conceived,* featuring characters *they invented,* set in worlds *they imagined.*
Now, many AI-driven tools help students skip this friction. Take what happens when you search “story ideas” on Google. The AI overview section at the top of the results page gets right into a bolded list of prompts like, “Imagine a celebrity on a family lunch, a mailman discarding letters, or a bank robber having an existential crisis.” Or, “A character discovers superpowers, a spaceship lands in a small town, or a dystopian society enforces strict uniforms” (Kay, n.d.). Without further conversation with a teacher or classmate, a stumped student can start drafting from these specific ideas. (Interestingly, as of this writing, any Google search for “story ideas” will result in this same list of specific prompts. Which may not come as a surprise to any ELA teachers who are starting to notice a lack of variety in students’ creative writing.)
A student looking for more tailored advice might go to ChatGPT and enter a description of the project. If they were to type, “I am confused about what to write for a multi-narrative story,” the model responds with a definition of the genre, followed by an example prompt: “A power outage affects an entire city—how do different people experience it?” Next, students are told which perspectives they could use in this story: “a nurse trying to get to a hospital, a teen running from a crime they witnessed, a lonely old man who sees this as a chance to reconnect with his estranged son.” The model then gives four options for how to structure the story, and finally, four more “story starters” in case the student doesn’t like the power outage idea. If I am a stumped student, it is unlikely that I will bypass this low-hanging fruit to tackle an original idea. And this, our ELA department thought, should prompt reflection from educators.
What is gained and what is lost by this new normal of students using AI prompting to break their writer’s block? In theory, students gain a personal writing coach—something that would seem like a “great equalizer.” Students also no longer face the embarrassment of seeking help from a classmate or the awkwardness of seeking it from a teacher. These gains—especially when we consider how AI can replace peer reviews—cement writing as a completely individual pursuit, through each of its steps. This, admittedly, made our department a little sad, as some of our best classroom experiences have been those “eureka moments” when students share breakthroughs during planning and editing.
A Question of Ownership
But the biggest loss comes with the AI prompting itself—particularly when models offer hyper-specific, ready-made story ideas. ELA teachers know there’s something special about students finishing narratives they conceived, featuring characters they invented, set in worlds they imagined. When students work with an original concept, they feel genuine ownership over both process and product. That ownership gives them the chance to think, It may not be perfect, but it’s mine.
Our department worried that when AI prompting becomes “step one,” only more privileged or gifted writers will experience that ownership, while struggling writers will default to building upon prepackaged stories. We wouldn’t “ban” AI tools—where could we even start? But we became vocal about how important it is for writers to own their story ideas.
We’ve begun celebrating original ideas before students start drafting, making it clear that authentic creative processes are messy and that our stumbles lead to our proudest successes. When students finish projects, we invite them to share how their ideas morphed as pieces took shape. The more open they are about this journey, the more willing their classmates are to replicate it.
