A few weeks ago, just days after I gave my AP Research students a stern but cautious speech about engaging with ChatGPT, College Board released a much-needed statement of guidance regarding the new variable of AI in the world of research and student writing. The tl;dr is this: Instead of ignoring it or forbidding it, we need to carefully educate our researchers regarding its ethical and unethical uses and teach them to think of AI as a writing tool the same as any other (except obviously infinitely more powerful and with the potential for abuse greater than anything they’ve seen before!).
I like this plan.
The Reality of Young Researchers
Without getting into the logistics of academic dishonesty and all the other moving parts here, I think this is the best practical and realistic choice College Board could have made. And I think AI is actually a fantastic tool. What I’ve discovered in my first two years of teaching AP Research is that our kids are really, really underequipped to explore the world of academic, peer reviewed research. At the risk of seeming mean, I think it’s worth considering why, and at the risk of being seen as lazy, I’m going to consider it using bullet points:
- They lack academic vocabulary. You don’t even need to go as far as specific fields of science to notice how limited their sense of sophisticated vocabulary is. This is a huge limitation when trying to gain a deep understanding of an unfamiliar concept.
- They have had limited practice with creative problem solving. Sorry English teachers, but we don’t do a great job of this! When I tell them they might have to “get creative” with search terms and combinations and to find some clever ways to find the information they need by playing around with diction and phrasing, most of them pretty much log off.
- They don’t know much about how academic research “works” yet. Consider the collected works we ask students to produce by, say, junior year that involve digging around in the true depths of academic journals or search engines. Do they know how to get there? Do they know when something is peer reviewed and why that matters? Do they understand why it’s useful to read beyond the “abstract”? You get the idea; the magnitude of information and unfamiliar structures here for them is enormous.
You Can’t Spell “VocabulAIry” without AI
I mean…you can. You actually HAVE to. But just leave my strained subtitles alone and hear me out here.
This is what my students are going to do this week as they get into the really difficult journal-scouring phase of their AP Research work. They’ve already got some fundamental knowledge of their fields, but they need to start considering connections and relationships between concepts. They also need to figure out what they DON’T know about their topic yet, which can be especially hard because…well, how do you search for unknown things when you don’t know where to look for them?
Here’s one way that’s pretty brand new and cool: You ask an AI to give you some words to help you look. For my model I am researching the merits of Standards Based Grading compared to a traditional 4 point scale approach. Here’s what ChatGPT generated when I asked it for a useful set of search terms to find good research in these fields (one of the nicest things about AI search generators is their ability to reply thoughtfully to pretty casually worded questions).

Nothing super fancy here, but notice what we might call the “supporting” language to try and poke at specifics. “Benefits of” “Implementation” and “Critiques” are all phrases our kids might not think of as search tools simply because they haven’t done this before. A great start!
But that looks pretty elementary if I’m being honest, and like I said, my kids are going into their second round of research so I think they’d already have a lot of the sources these phrases might generate.
Swift Sorting
I know that sometimes my students struggle to sort all the tidal waves of results they get from plugging ANYTHING into a search engine, so I asked ChatGPT to chunk these results into some more useful categories. When my students see results like this, they can start to think about the usefulness of each one to their current work, discard what isn’t that useful, and consider whether some represent a direct line to what they really want to know.

Filling in the Gaps
My final “ask” of ChatGPT is the one that I’m going to ask my kids to do the most work with. It’s half practical research move, half reality check. I instructed the AI to give me a list of “other important terms I should know” that related to my broader concepts. As an educator who already knows all sorts of stuff about my topic, I know that this is an excellent list that would send an actual inexperienced researcher in useful directions:

This is only one section of a list seven categories long that the AI generated for me! I’m going to ask my kids to take this final list for their topics and sort it into columns for themselves: Terms I’ve heard or seen, terms that are brand new, terms that I know but haven’t tried exploring yet. They can discard the ones that they’ve already worked with.
In a course where I have 40 researchers pursuing just as many independently chosen topics, this eliminates an entire slow and painful process for us collectively. I don’t have to help students vocab-build individually through exhaustive seek-and-find searches of their initial research findings (which, by the way, can lead to all sorts of dead-ends because they struggle to know which terms are likely to yield something useful), and they don’t have to rely on me to help them know all of the jargon they don’t know to look for.
So there you have it, one really practical use of AI that’s safe and ethical and promising. Be cautious with your student researchers before sharing this with them; they have to know that it doesn’t take too many keystrokes to ask an AI something that would cross the line. But with some early guidance, this activity can help them have really interesting, highly relevant sources in their hands days sooner than they would if they had to trudge through the entire process on their own.
What do you see as the role of AI in research? Elsewhere in writing? Let me know on Twitter @ZigThinks
–Mike
