Every psychology professor knows the feeling: a mountain of readings, a blank slide deck, and a Sunday night deadline. This is the workflow I actually use — not a theoretical one — to go from raw research to a lecture I feel good about.
I start by feeding AI tools the source material, not asking them to invent content from memory. That single habit prevents most of the accuracy problems people run into.
Here's the actual first prompt I use, almost word for word: "Here are three research articles on [topic]. Read them closely and pull out the three ideas a student would most need to understand before an exam. Don't add outside information — only work from what's in these articles." That one instruction — don't add outside information — is doing almost all of the work. It keeps the tool anchored to real, citable research instead of quietly blending in things it half-remembers from training.
From there, I ask for a structured outline organized around a single question students should be able to answer by the end of class, then I build examples and discussion prompts on top of that skeleton myself.
For a lecture on classical conditioning, for instance, that single question might be: "Why does a fear response generalize to situations that look nothing like the original trigger?" Once I have that anchor question, I ask for three different real-world examples that illustrate it — not because I can't think of examples myself, but because reviewing three drafts, even mediocre ones, is faster than staring at a blank page waiting for the right one to appear.
Before anything goes on a slide, I check every claim against the original article myself. This isn't optional — I've caught AI summaries that flattened an important caveat or attributed a finding to the wrong study. The tool drafts the scaffolding; my own expertise still decides what's actually true.
The result isn't a shortcut around teaching — it's a shortcut around the blank page, which is the part that actually eats your evening.
If you want to try this yourself, start small: pick one dense article you already know well, and ask an AI tool to summarize only what's in it, nothing more. You'll learn more about where these tools help — and where they don't — from that one exercise than from reading ten articles about "AI in education."