A worksheet gets photographed at 9:40 on a school night, and the app hands back the finished answer. A grandmother discovers that her clinic's portal now has an AI assistant that asks how she is feeling, and she finds herself liking the exchange.
Neither of those is wrong on its face. AI can be a terrific tutor. It can explain the same concept for the fifth time without getting impatient or generate extra practice problems. I like AI, and I use it because it is useful, which is exactly why I think the small choices around it matter. They are made when someone is tired, stuck, lonely, or running late, and by the time anyone looks up, a habit is already in place.
That is why the home matters so much here. A child picks up a working theology of technology long before she could put one into words, from what her parents praise, what they sigh at, and what they put face-down on the table when she walks in.
Even the Honest Work Gets Suspected
Recently my son showed me part of his homework and, before I had read any of it, told me it sounded like AI. He was right that it did not sound much like him, because he does not write the way he talks. Then he added, "I did not use AI, though. I use my brain."
Nobody had accused him, and that was what stayed with me. He already knew that writing can now be judged on whether it looks too polished, and that the more carefully a student writes, the more suspicious it can look. The shortcut is the danger most of us see coming. The quieter one is a home, or a classroom, where honest effort has to keep explaining itself. We should still ask questions about AI, but perhaps in a way that assumes the best about the child in front of us until there is a reason not to.
Questions That Travel
Rules are a natural first move, and some of them are good. "Devices sleep in the kitchen" solves several problems at once. But a rule written for last year's technology may stop fitting this year's, and a rule with no reason attached can raise children who are very good at staying just inside the line and less good at understanding what the line was there to protect.
My coauthor, John Delano, and I have been working on a book about navigating AI with biblical wisdom, and one idea I keep returning to is that good questions travel better than rules. They can follow a family from the homework spread across the table to the phone under the pillow.
With homework, the question I care about is whether the help produced understanding or only an answer. A child who is embarrassed to ask her teacher to explain something one more time can ask AI as many times as she needs, and I love that use of it. The test comes afterward, when I ask her to walk me through it without the tool. If she can, the tool helped her learn. If she cannot, it may have given her an answer without helping her understand it.
Loneliness is harder. A teenager may turn to a chatbot because it is always available, easy to talk to, and does not feel awkward in the way a real conversation sometimes can. The grandmother who likes being asked how she is feeling is responding to something similar, and it is worth taking seriously how many people are. A 2025 national survey found that nearly three in four teens had tried AI companions, and among those who used them, about a third had chosen one over a person for a serious conversation (Robb & Mann, 2025). A recent review in Pediatrics adds that younger children in particular may struggle to separate a responsive system from a person (Grundmeier et al., 2026). That does not mean the answer is simply to take it away. It may be more important to ask why the chatbot has become appealing in the first place, and whether there is a real relationship that needs more attention. One line is worth saying at home, and saying more than once: AI can be an extraordinarily useful tool, and it is not a person.
And then there is the question I find most uncomfortable, because it turns back toward the adults: would I want my child copying what I am doing right now? Children notice when the parent who bans AI shortcuts has AI write the thank-you note to their teacher, but the mirror works the other way as well. A parent who says out loud, "That sounds convincing. How could we check whether it's true?" or "I could have AI write this message, but this one needs to come from me," is teaching children how we decide when to use it. Most families will get some of this wrong, and I know we will. A parent who can say, "I have been rushing through things that deserve more care, and I want to change that," may teach something a perfect record never could.
What This Looks Like on a Tuesday
The best family AI education may happen when no one announces that it is happening. A child asks why the moon looks enormous near the horizon, and instead of simply answering, her mother pulls up AI, asks the question with her, and then asks how they would know if the answer is right. On a road trip, a family has AI invent trivia about the states they are driving through, someone catches it getting a fact wrong, and now the children know a confident answer is not necessarily a correct one.
Common Sense Media recommends exactly this: exploring these tools alongside children and deliberately testing them on things the family already knows, so that kids learn a useful tool can still be questioned (Common Sense Media, 2025). It is an old pattern: Israel was told to teach its children when sitting at home and when walking along the road, when lying down and when getting up (English Standard Version Bible, 2001/2016, Deuteronomy 6:7). Nothing in that verse waits for a scheduled lesson.
The Crooked Dragon
Most of these questions are about how to use AI well. I have also been wondering what is worth doing without it, and the refrigerator door keeps coming to mind. An image generator can turn out a spectacular dragon in six seconds, and making one together can be fun. Meanwhile, the crooked dragon taped to the refrigerator took a child a whole afternoon. She had an idea, worked it onto paper, gave the dragon purple wings and seven toes, and decided it was worth showing somebody. Something happened in her over that afternoon that no generated image would have given her, and I hope families keep making room for it, even when a tool could finish the job faster.
Sources
Common Sense Media. (2025, July 17). Parents' ultimate guide to generative AI. https://www.commonsensemedia.org/articles/parents-ultimate-guide-to-generative-ai
English Standard Version Bible. (2016). Crossway. (Original work published 2001)
Grundmeier, R. W., Fiks, A. G., Jenssen, B. P., Proctor, S. N., Ferro, D. F., & Johnson, K. B. (2026). Generative artificial intelligence: Implications for families and pediatricians. Pediatrics, 157(4), e2025074912. https://doi.org/10.1542/peds.2025-074912
Robb, M. B., & Mann, S. (2025). Talk, trust, and trade-offs: How and why teens use AI companions. Common Sense Media. https://www.commonsensemedia.org/research/talk-trust-and-trade-offs-how-and-why-teens-use-ai-companions