← Home / Approach
What I believe
We can’t prohibit our way out of this.
Banning AI without thinking through the effect on students and educators doesn’t protect learning. It just moves the conversation somewhere I can’t reach it. These tools are widely available, and pretending otherwise leaves everyone unprepared. The work is figuring out where AI genuinely helps, where it quietly harms, and how to tell the difference.
The most important thing is that we stay the human in the loop: consumers of AI output, using our critical digital literacy skills to decide when, where, and how to use these tools.
It’s a tool, not intelligence
Generative AI isn’t artificial and it isn’t intelligent: it’s computer programming that predicts the next word or pixel. John Warner, author of More Than Words, calls it a “syntax generator,” which is a more honest name than the one we’re stuck with. Naming it accurately is the first step to teaching with it well.
Pedagogy comes first
AI is not an add‑on. Start with the student learning outcomes you care about and the pedagogy you believe in, then ask whether a tool serves them.
The science of learning still applies
Retrieval practice, low‑stakes assessment, and timely feedback are what move learning. AI is useful precisely because it makes those things faster to build.
Ethical and legal use is the floor
Privacy, transparency, and honest attribution aren’t afterthoughts. Policy has to come before adoption, not after the first incident.
Skeptics are welcome
AI is not one size fits all. Discipline, context, and purpose all matter. Choosing not to use AI is legitimate: you just need a real rationale, and a willingness to hear other ideas.
Metacognition over enforcement
I’d rather students reflect honestly on what a tool did well and badly than play cat and mouse with detection software that doesn’t work.
You don’t have to be the most curious person in the room
You just have to find your people, the ones already playing with these tools, so you can stay current without making it your whole personality.
Frameworks I use
Good policy is assignment‑level, not building‑level.
A blanket rule at the top of a syllabus tells students almost nothing. Guidance has to live where the work does, on each individual assignment. Click any of these to read more.
AI Stoplights: assignment-level permission, at a glance
A red, yellow, or green signal on every assignment so students never have to guess what’s permitted. Red means no AI. Green means use it freely. Yellow means specific, stated uses only.
It takes minutes to apply and removes nearly every ambiguity that leads to an academic integrity conversation. It also gives students and instructors a shared vocabulary, which matters more than any single rule.
- Free to adapt for your own courses
- Works at course, department, or program level
- Students stop guessing; instructors stop policing
AI Acknowledgement Statements: reflection, not surveillance
On major assignments, students fill in a short sentence frame: which tool they used, what it did well, what they had to change and why, and whether they’d use it again. Students who chose not to use AI have their own frame, and it counts equally.
The point is metacognition. Naming what you had to fix is where the critical evaluation happens: you can’t answer it without having actually judged the output. The simpler the frame, the better it works.
My full course AI policy: free to adapt
The actual AI use policy from my syllabus, published in full: what counts as an AI tool, permitted uses, where AI can’t do the thinking, FERPA and data privacy, academic integrity, APA citation guidance, and both acknowledgment templates.
I drafted it, revised it with Claude for 25–26, and updated it again for 26–27. Take it and make it yours.