The AI We Use and the Literacy We’re Missing

If you ask a Colby College student whether they use ChatGPT, you’ll usually get one of three responses:

1. an immediate “Of course not” (delivered a little too quickly), 
2. a polite “only for brainstorming,” or 
3. a quiet nod, as if admitting to a minor crime. 

The irony is that nearly everyone is using AI, and most of us are using it far more confidently than we probably should. That’s exactly why the College needs campus-wide AI literacy: not because students are doing something wrong, but because we’re navigating a technology we were never actually taught to understand. 

ChatGPT didn’t arrive slowly. It landed on campus the way a Maine winter does—suddenly, and encountering wildly different levels of preparedness. Within months, AI became part of job applications, internship tasks, research projects, and the daily academic shortcuts that no one ever admits out loud. Yet students’ understanding of AI still depends almost entirely on their major, their professors, and whether someone happened to forward them an email about a workshop. Some students get structured exposure, but many never receive any formal instruction at all. Most of us end up improvising.

Improvisation works fine when you’re deciding how much dining hall food to risk in a single sitting. It’s worse when employers now assume new graduates already know how to use AI responsibly, efficiently, and with more nuance than “paste the prompt and hope for the best.” Generative AI feels intuitive. You type, it answers—but that illusion of simplicity hides the real complexity. Writing an email with AI is not the same as recognizing when it’s fabricating information, spotting subtle bias in its output, avoiding unsafe data practices, or evaluating whether the answer actually makes sense. Without guidance, students end up learning through trial, error, and the occasional moment of panic when AI confidently invents something that absolutely never happened. 

The heart of the issue is that AI literacy at the College is uneven. Students in computational fields tend to encounter AI as part of their coursework, while students in the humanities or arts might graduate without ever discussing how AI intersects with their discipline at all. And even within departments, faculty vary widely in how they handle generative tools. One professor might encourage thoughtful experimentation; another might ban AI entirely; another might permit certain uses but never explain the limits. When norms shift from one classroom to the next, students are left guessing what responsible use actually looks like or avoiding the technology altogether out of fear of messing up. 

The College isn’t behind compared to peer institutions, but we are at a transition point. The Davis Institute for AI is the first cross-disciplinary AI institute at a liberal arts college in the country, and it runs workshops, demos, discussions, and faculty collaborations that many institutions would envy. But even with this strong foundation, the gap remains between the Institute’s expertise and the everyday reality of student learning. Students who already feel comfortable with technology show up to events; students who don’t often never find their way in. And while some classes incorporate AI meaningfully, others understandably have no framework yet. 

Strengthening AI literacy at the College does not require reinventing the curriculum. It requires weaving AI into the places where students already learn. One practical approach is to integrate short, discipline-specific AI modules into existing courses. A sociology class might explore how AI interprets qualitative data. A biology class might analyze AI in protein modeling. An English class might critique AI-generated interpretations. Students would begin seeing AI not as a mysterious external force, but as a tool that intersects differently with every field. 

Clarity is equally important. Right now, students navigate a maze of mixed messages about what counts as acceptable use. A simple campus-wide set of guidelines would help everyone understand when AI enhances learning and when it compromises it. Students shouldn’t have to guess whether something is allowed or whether a model is safe to put personal or confidential information into. Guidance makes responsible use possible.

And, of course, none of this works without faculty support. Instructors can’t be expected to guide students through AI if they haven’t been given the time or training to navigate it themselves. The College already has promising structures: FIT Fellowships where faculty rethink course design with generative AI in mind, Haynesville support for research, informal faculty discussions, and exposure to privacy-preserving tools like NotebookLM. Expanding these resources into flexible, structured training would help faculty integrate AI at a pace that respects their existing workloads while still pushing the curriculum forward. 

The first step, then, is supporting instructors. When faculty feel prepared, everything else becomes possible. Students receive clearer expectations, modules become easier to adopt, and the broader campus culture shifts from uncertainty to confidence.

Preparing students for an AI-mediated world is not a departure from the College’s liberal arts mission, but the next logical step in it. AI is now part of research, communication, creativity, policy, environmental work, data analysis, and nearly every field our graduates enter. A poet, a physicist, and a political scientist will not use AI in the same way, but each deserves the literacy to understand its possibilities and its limitations. 

Everyone uses AI. It’s time we learned how to use it well.

 

~ Aaminah Ali  `29