Grammar, Gatekeeping, and Generative AI: A Case Study from My Kitchen Table
I have somehow managed to become a fully functional adult: former teacher, author, and novelist, and parent – without ever truly understanding the difference between a compound and a complex sentence.
There, I said it. { Looking at you, SB :) }
Grammar has always been one of those things I’ve bristled at being taught in isolation. I know the chair of my daughter’s English department (not SB, but another dear friend) – she’s a ***phenomenal*** educator – and yet this is the topic where we’ve been known to spar a little.
I’ll admit, I’ve teased my English teacher friends that they treat grammar as if it were carried down from Mount Sinai on stone tablets. But the truth is, grammar evolves. It changes with time, with context, and with audience. What counts as “good grammar” in one space is sometimes just a gatekeeper’s code in another.
That said, the point of this post isn’t a manifesto on grammar instruction. It’s about how I – a skeptical, middle-aged former educator – ended up using AI as a grammar tutor, and what that experience revealed about how parents, teachers, and students can use these tools productively.
The Experiment
Here’s how it started:
Last night, my 8th grader asked me for help with her upcoming Grammar Test. {Hello SM… :) } Usually it’s on a topic I am familiar with, or even excel at talking about. For example, even after nearly 40 years, I still have a strong grasp of 8th-grade algebra. And in the same English class, they’re reading Lord of the Flies, one of my all-time favorites. It was a unit on sentence types, which, candidly, I couldn’t identify – simple, compound, complex, and compound-complex – if they were already labeled. So I decided to use ChatGPT to create a practice quiz for my daughter.
“Create examples of simple, compound, complex, and compound-complex sentences – make them about Lord of the Flies.”
Instantly, it generated categorized examples about the book. So while they’re currently reading that novel, to my understanding, the two units are not connected. Hmmf.
Because my prompt wasn’t tight or specific, it created 20 sentences and labeled them. Turns out I can identify them when labeled, which still didn’t mean much. I mean, ChatGPT did exactly what I asked for – which is the trick about AI: you have to pretend you’re talking to a 3-year-old - literal, concrete.
Immediately seeing my error, I quickly generated a revised version of the prompt, explaining I wanted to make a quiz for my daughter. It generated a list of another 20 sentences, and she was to respond 1, 2, 3, and 4 for each of the different types of sentences.
Her first score? 50%.
So I asked:
“Which sentence types did she struggle with? Give me 10 sentences of those types. Let’s try again.”
The AI diagnosed her weak spots (compound and compound-complex, apparently) and built a new set. She took the quiz again – and this time scored 80%.
That’s when it hit me: I had just experienced AI as an adaptive tutor. It wasn’t doing the work for me; it was doing what great teachers do – diagnosing, scaffolding, and adjusting to my daughter’s needs.
What AI Actually Did
Under the hood, here’s what happened:
Generated content aligned to a literary text (Lord of the Flies).
Assessed her understanding through a self-created quiz.
Provided feedback and explanations for each incorrect answer.
Adapted instruction to target her weaker areas.
Tracked progress from 50% to 80%.
Encouraged reflection (“Which types does she struggle with?”)
It was immediate, relevant, and – importantly – responsive.
No grading lag. No red pen. No shame. Just learning through iteration. Now, that’s #LifeReady.
And, most surprisingly of all, it was something I could do as a parent. I didn’t need to remember the definitions of sentence types or even follow every reference to Lord of the Flies. The AI handled the content; I handled the curiosity.
That’s a small but profound shift for those of us who grew up in an analog classroom. Parents of a certain generation know the feeling – the way we learned to “solve for X” isn’t the way it's taught now. The vocabulary, the process, and even the philosophy of learning have changed. But with tools like this, I could reenter my kid’s learning world without pretending to be the expert. I could model something more useful: how to ask better questions, how to learn alongside her, and how to see technology as a bridge – not a barrier – between generations.
But what if I wasn’t handling the curiosity — it was my daughter? Don’t tell her that, because, secretly, “helping” her with her HW is singularly the best part of my day.
The Real Lesson: Clarity Over Cleverness
What made the experience powerful wasn’t the AI itself – it was the clarity of the prompts.
Each time I gave specific directions – “Make them about Lord of the Flies,” “Include explanations,” “Identify her weak spots” – the AI delivered precise, valuable results.
Vague directions produced vague answers.
Clear directions produced personalized learning.
That’s the literacy skill I wish every student – and teacher – learned early:
Prompting is a form of instructional design.
It’s the 21st-century version of asking better questions, setting clear objectives, and thinking critically about what you need from a tool. In other words, prompting is metacognition in action – a digital mirror for how clearly we think. The skill isn’t in manipulating the algorithm; it’s in articulating intent. When you know what you’re asking for, the AI becomes a collaborator. When you don’t, it just reflects the fog back at you. The same principle that makes a good student inquiry or a good research question now applies to every interaction with technology: clarity of purpose leads to quality of insight.
AI, Kids, and “Cheating”
For what it’s worth, neither of my kids is racing to use AI.
Both my college sophomore and my 8th grader think it’s cheating. Despite all the headlines about students using it “rampantly,” I’ve found that’s just not true – not because of fear of being caught, but because it assumes a level of tech proficiency and confidence not all students have.
More importantly, it assumes that kids want to take shortcuts. Most don’t.
When they do use AI, it’s usually when someone has helped them see how it can support their thinking – not replace it. That’s what this little grammar experiment reminded me of.
What This Means for Classrooms
AI can:
Generate practice material tied to current readings.
Provide instant, low-stakes feedback.
Differentiate instruction based on error patterns.
Reinforce self-assessment and reflection.
Free teachers from repetitive grading tasks.
But it all depends on how we teach students to use it.
We don’t need to outlaw it — we need to coach them to collaborate with it. I’ve written about this often, here and on LinkedIn: the real danger isn’t students using AI; it’s students not learning how to use it well.
Take my own kids. Both are thoughtful, capable students — one in middle school, one in college — and both see AI as an academic taboo. My son, for instance, is on the Dean’s List at a top university, majoring in one of the most challenging fields imaginable. Yet he refuses to feed his semester’s worth of calculus or physics notes into an AI to generate practice questions or summaries for finals. I’ve suggested it countless times, but he won’t do it. He’s internalized what a lot of students have: the idea that using AI equals cheating, even when it could be the single best study partner he’ll ever have.
And that, to me, is the missed opportunity. Imagine if schools didn’t just warn students about plagiarism, but actually taught them how to think with these tools. How to prompt, test, refine, and verify. We’d move from a culture of fear and prohibition to one of curiosity and fluency. AI isn’t replacing learning; it’s reshaping what learning looks like. The question isn’t whether students should use it — it’s whether we'll teach them to use it wisely.
Results and Reflection
And while I still couldn’t diagram a sentence to save my life — seriously, is that still taught? It has to be one of the clearest examples of clinging to outdated skills in a 21st-century classroom. Not to belabor the point, but I have two master’s degrees from world-class universities and haven’t earned a grade below a B since the first Bush administration — and not once, since seventh grade, has anyone asked me to diagram a sentence. It’s a relic of a time when we thought labeling parts of speech was the same as understanding how ideas fit together.
Last night’s homework illuminated a more significant understanding:
AI can make learning more human when it’s used for curiosity, not compliance.
It’s not about the grammar. It’s about the thinking.
So yes, maybe my daughter’s English teacher wins this round – clarity still matters, even when the rules keep rewriting themselves.
Whether it’s grammar or generative AI, the real lesson is the same: clear thinking leads to clear expression.
Good writing isn’t about obeying every rule; it’s about communicating meaning with intention. And when we teach kids – or ourselves – to ask sharper questions, to give better directions, to name what we need from a tool or a sentence, we’re really just practicing clarity in thought.
The grammar rules may change. The technology certainly will. But clarity – that’s the part that endures.
Discussion Prompt
If you’ve used AI in your classroom or home, what surprised you most – the technology, or the way you (or your students) adapted to it?