The AI Read My Kid’s Essay Before I Did
I was waiting outside the classroom for a parent-teacher conference when a fourth-grader named Marcus walked up to me, holding a crumpled worksheet. His mother was still on her phone, one earbud in, nodding along to something I couldn’t hear.
“She already knows what you’re going to say,” Marcus whispered. “She asked the AI to predict my grades.”
I laughed, thinking he was joking. Then his mother looked up and said, “The chatbot said you’d probably recommend reading intervention. Should I do that, or is it fine?”
I didn’t know what to say. She wasn’t being rude—she was actually trying to be efficient. But Marcus was standing right there, waiting for her to look at his paper. She never did.
That was the first time I realized that “screen time” no longer describes what’s happening in families. It’s not just scrolling or playing games. Many parents are now outsourcing large parts of their thinking to AI: homework help, meal planning, even conversations about how their child is feeling.
I’ve started asking students about this in class discussions, mostly because they bring it up first. One girl told me that when she fights with her little brother, her dad holds up his phone and says, “Let’s see what the AI recommends for sibling conflict.” Another boy said his mom sends his report card to a chatbot and reads the analysis out loud, but she hasn’t actually read his test questions.
At first I felt frustrated. Not at the parents exactly—I understood the exhaustion. Many of them are stretched thin, working late, trying to keep multiple kids alive. AI offers a shortcut to answers and reassurance. But the kids are noticing something important: the person who used to be their sounding board is now often a voice from a speaker or a text bubble.
The biggest shift isn’t what the AI says. It’s that the parent is no longer fully present in the moment of uncertainty. When a child says, “I don’t know if my project is good enough,” the old response used to be a parent leaning over the paper, frowning, asking questions. Now the response is often, “Let’s run it through here and see what it thinks.” That changes the emotional experience for the child. Instead of being met by a parent’s attention, they’re being met by a tool. And tools don’t care about them.
I saw this play out one afternoon in my own classroom. A student, Maya, had written a heartfelt paragraph about her grandmother’s cooking. Her father had apparently pasted the draft into a GPT app and asked for “suggestions to make it more vivid.” He then gave Maya back a list of edits that included the word “succulent” and a metaphor about time swirling in the kitchen pot. Maya told me, “He thinks that’s what I should write, but I just wanted to tell the story.”
We spent a while talking about her paragraph, and eventually she kept her original version. But that moment taught me something practical: when parents use AI to “improve” their child’s work, the child often feels that their own voice isn’t worth listening to. They don’t hear it as help. They hear it as, “You’re not good enough as you are.”
Now, I don’t think AI is evil. I use it myself. I’ve generated lesson plan options and asked it to brainstorm alternative ways to explain fractions. But I’ve seen a pattern emerging with parents who lean on it constantly. They’re losing the small, messy interactions where empathy and judgment are built.
There’s a particular phrase I keep hearing from parents: “I just ask the AI for a second opinion.” That sounds harmless. But when a child asks “Why do I have to do this?” and the parent responds with “The AI says it’s important for brain development,” the kid learns that their parent isn’t willing to bend. The AI becomes the authority. The parent becomes the messenger.
I’m not the only one noticing this. A school counselor friend told me she’s had more children say things like, “My mom talks to her phone more than me.” She’s started calling it “substitute parenting” in her mind—not because the parents don’t care, but because they’ve found a tool that feels like it cares back, instantly, without exhaustion.
The tricky part is that AI responses are often soothing. They validate the parent. When a mother worries about whether she’s too strict, the chatbot says, “It’s okay to set boundaries.” That feels like support. But the child in the next room hears her reading that out loud and thinks, “She just learned about boundaries from a chat screen.”
So what works? I don’t have a perfect answer, but I’ve seen things that help in small ways.
One classroom parent decided to create a rule: no screens at the dinner table, including hers, even when she was tempted to ask the AI for a conversation starter. She told me she now uses a simple notebook titled “Questions for Kids” and writes down one silly question on her commute home. Her children are more engaged because the questions are weirder than anything AI generates—things like “If the moon was made of cheese, would mice move there?” That’s not a big change, but it gives kids the feeling that their parent is thinking about them specifically.
Another parent realized she was always turning to AI for homework help. She changed course: when her daughter asked for help, she would say, “Okay, let’s struggle for five minutes first. Just us.” After that, she allowed the AI to serve as a fact-checker, not a first responder. Her daughter started feeling more confident because the parent’s uncertainty was present, and it was okay.
For teachers, I’ve learned to ask parents a gentle question during conferences: “What did you notice when you read your child’s draft before or after the AI?” If they say, “I didn’t read it, the AI did,” that’s a signal. I don’t judge. I just model what I do with my own son. I show him my corrections in handwriting and say, “I wrote this part wrong too, look.” He likes seeing my mistakes. It makes him think I’m real.
I’ve also started saying, in class, “I want your thinking, not ChatGPT’s thinking.” It sounds blunt, but kids get it. They actually look relieved. One boy told me that when his dad pasted his science project into an AI and asked for a “more neutral tone,” his dad lost interest in the actual experiment. We talked about that. He eventually redid the conclusion with his own voice, and his dad only saw the final product after it was printed. That felt like a win to me, even if it took a week.
I still slip up. I catch myself scrolling for advice on my phone instead of just staring at the ceiling during a confusing moment in class. And I know parents who are on their phones are often thinking about their kids—not ignoring them. But the kids are more demanding than the AI. They want a live reaction. They want a parent who stumbles, says “I don’t know,” and then picks up a pencil or sits on the floor.
At the end of the school year, Marcus came to my desk and showed me his reading log. His mom had written a tiny note in the margin: “You read this well. I watched you.” That was all. No AI summary, no printed analysis. Just three sentences in her own messy handwriting.
He smiled and said, “I think she’s getting better at reading with me.”
I hope she is. I think we all are. Managing our AI use isn’t about going back to the old days. It’s about remembering that when a child says “Look at this,” they don’t need a perfect response. They just need us to look up.
Please indicate: Thinking In Educating » The AI Read My Kid’s Essay Before I Did