Our students are going to be expected to use AI — in their majors, their jobs, their civic lives — whether or not any of us teach them how. That makes engaging with AI in higher ed less a trend to chase or resist, and more a responsibility: we owe students the judgment to know when AI helps, when it doesn’t, and how to be honest about which is which.
The AI Sketchbook is where faculty, librarians, technologists, and students compare notes on what that looks like in practice — a working record of what’s been tried, what’s worked, what hasn’t, and what’s still unresolved.
The question isn’t whether students will use AI. It’s whether they’ll learn to use it with judgment — knowing when to reach for it, when to set it aside, and how to say plainly what they did. That’s not a skill students pick up on their own; it’s one we have to teach, the same way we teach foundational knowledge, source evaluation or research methodology.
It’s worth naming why that’s hard. AI makes academic dishonesty easier and can quietly hollow out the low-stakes, formative work — reflective drafts, early problem sets — that helps students find their own confusion. It fabricates sources with unearned confidence, reflects the biases of training data that is disproportionately English-language, Western, recent, and already digitized, and produces fluent, well-structured prose that can be subtly or badly wrong.
None of that is a reason to disengage — it’s the reason engagement has to be deliberate. The difficulty of working through an argument, writing your way toward an idea, testing evidence, or confronting sources and data that resist easy interpretation: these are key learning moments.
Students already sense this. Many use AI to save time, but do the work themselves when they know how it actually matters to them. But they’re not always sure where that line is, or how to talk about it openly. The typical assessment regimen of higher education make skipping the friction feel rational, which is exactly where guidance from faculty, librarians, instructional designers, and technologists matters.
No single discipline owns these questions. Humanists have long studied interpretation and authorship. Librarians bring deep expertise in information literacy, metadata, access, and source evaluation. Social scientists bring methods for studying data, behavior, and inequality. Technologists understand systems and infrastructure and access issues. STEM faculty and students run into still other parts of the problem, in labs and problem sets where AI raises its own questions about process and rigor.
These are not peripheral concerns. They are central to the future of AI use, which will be in the hands of our students. The skills that make AI use meaningful rather than mechanical — evaluating sources, recognizing bias, attending to what’s missing, understanding method, documenting process, judging consequences — are distributed across the university. That’s the case for sharing what we learn instead of each of us working it out alone.
Many perspectives, many partners
The sketchbook does not claim AI as the territory of any one field or office. It offers concrete examples that can travel across units, disciplines, and conversations. We hope to faciliate a broad community of practice
Teaching sketches focus on assignments and classroom setups where AI becomes an object of critical inquiry — situations where using AI teaches students something about how knowledge gets made, evaluated, and trusted. The goal is never to outsource thinking; it’s to make the thinking more visible.
Policy sketches offer adaptable language and decision tools for syllabi, assignments, programs, and departments. They are not model university policies. They are prompts for making local choices visible: what AI is allowed to do, what students must still demonstrate, and help students make responsible AI choices.
Research sketches document workflow ideas: what it actually took to bulk-process a set of archival documents, build a map from a folder of photographs, or generate a 3D model from a line drawing. They try to provide a balance of limitations and possibilities
Sketches are tagged by status — rough, tested, or refined — so you can tell what’s been tried once versus what’s been iterated and classroom-tested.
Read for adaptation
The sketches are meant to be useful even as works-in-progress. The caveats, failures, and local constraints are often the most transferable parts.
Each sketch documents a specific experiment, framework, or policy idea: a single assignment, a research workflow, a tool used for a defined purpose, or a piece of language that helps clarify expectations. They share enough structure to make them easy to evaluate and adapt, but they are not meant to be formulaic.
They are not refined advice or best practices, although some may become so over time. Any ideas that you think are worth trying can probably help someone else.
We owe it to our students to figure this out, and that’s a lot easier together, from various vantage points, than alone. If you’ve tried something with AI in a class or research project — and you have something honest to say about how it went — we want to hear about it. Rough accounts and failed experiments are exactly what this site is for.
The easiest way to start: email us a draft. A few paragraphs describing what you tried and what happened is enough. We can help shape it into a sketch.
If you want to submit directly and own your sketch from the start, the contribute page walks through that process too.
Or drop by Amaranth studio hours at Mesa Vista Hall 2068.
The AI Sketchbook is a project of Amaranth Digital Humanities Studio at the University of New Mexico, built with the Xanthan open-source framework for academic Jekyll sites.