What 'Alongside a Parent' Looks Like When the Tool Makes Things, Not Just Answers Questions

This post cites Common Sense Media's published guidance for context. Common Sense Media has not reviewed, endorsed, or partnered with Xyplor — the citation is attribution, not affiliation.

TL;DR: Common Sense Media's "AI Chatbots and Your Child" guidance (June 2026) recommends no chatbots at all for kids 5 and under, "limited use, alongside a parent or caregiver" for ages 6–12, and supervised-but-not-for-comfort use at 13+. The guidance is written about chatbots — tools built around asking and answering. This post walks through what "alongside" actually looks like when the tool isn't a chatbot but a maker: a kid building a game, a podcast, or a storybook. The short version: alongside means visible, not necessarily present. Here's the distinction, and the question Common Sense suggests teaching that we think is worth adopting regardless of which tool a kid is using.

What Common Sense Media actually says

Common Sense Media published age-and-stage guidance on AI chatbots in June 2026. It's worth reading in full, but the headline recommendations are specific:

5 and under: no AI chatbots. Young children can't reliably tell AI apart from a real person.
6–12: "Limited use, alongside a parent or caregiver."
13+: supervised use is fine for homework and creativity — not for comfort or feelings.
All ages under 18: no AI for emotional support, mental health advice, or "friendship," full stop.

That last line is unambiguous and we think it's correct for any product serving kids, including ours.

The guidance was written about chatbots — Replika-style companions, general assistants, homework-help bots. That's the right scope for Common Sense to address first; chatbot companionship for kids is a live and growing concern. But a lot of parents reading it are trying to apply "alongside" to a different kind of tool, and the fit isn't exact.

The guidance is about chatbots. A maker is a different shape of tool.

A chatbot's whole interaction is a conversation. The kid asks, the AI answers, the kid asks again. "Alongside a parent" in that context reads naturally as present for the conversation — because the risk is in the conversation itself. If a bot starts role-playing as a friend, or a kid starts asking it for comfort instead of a parent, you want an adult close enough to notice in real time.

A maker tool is a different shape. A kid describes what they want to build — a platformer where the dragons need to be faster, a podcast about their favorite soccer team, a quiz for their little brother — and the AI produces something the kid then evaluates, keeps, or throws out and tries again. The conversation is in service of an output. The risk profile shifts from "what is the AI saying to my kid" to "what is my kid asking for, what did the AI build, and is any of it going anywhere public."

That's not a smaller risk. It's a different one. And it means "alongside" can be satisfied by design, not just by physical presence.

Why the distinction matters practically

If you read "alongside a parent" as "I need to sit next to my 9-year-old for every session," that's not sustainable for most families, and it's not really what the underlying concern requires for a creation tool. What the concern actually requires is:

You can see what your kid asked the AI for.
You can see what the AI produced.
Nothing leaves the house (gets published, shared, sent to a stranger) without you knowing.
The AI isn't positioned as a friend, confidant, or source of comfort — ever, for any age.

Those four things can be built into the product instead of depending entirely on a parent's physical attention span on a Tuesday evening. That's the design bet we made with Xyplor, and it's worth naming plainly: it's a bet, not a guarantee, and it doesn't make the chatbot-specific risks irrelevant — a kid can still ask a maker's AI something off-topic, which is why the same safety layers apply across the whole product, not just the "making" part.

What "alongside" looks like for a creation tool

Concretely, here's what we think satisfies the spirit of "alongside" when the tool is for making things rather than asking questions.

Visible prompts, not a locked diary

Every prompt a kid sends and everything the AI builds in response should be visible to a parent — not buried in a settings menu, not requiring a support ticket, just there in a dashboard a parent can open anytime. In Xyplor, every AI conversation is logged and visible in the parent dashboard by default. That's the closest digital equivalent to "alongside": you weren't in the room, but you can see exactly what happened, as often as you want to look.

Parent approval before anything goes public

A kid building something is different from a kid publishing something. The first is private practice; the second is a kid's creative work going in front of other people, including other kids. Xyplor requires parent approval before any creation can be published to the public gallery. A kid can build, iterate, and rebuild as much as they want privately — the approval gate only sits between "I made this" and "other people can see this."

The kid stays in the director's seat

This one is less about safety and more about what the guidance's underlying worry actually is: a tool that does the thinking for the kid versus a tool the kid directs. A chatbot that answers homework questions risks becoming a tool that finishes things for a kid instead of helping them learn. A maker has the same risk in a different costume — a kid can type one vague prompt, accept whatever comes out, and never touch it again. The alternative is a kid who gives feedback ("the boss is too easy," "I want the podcast intro shorter"), evaluates what came back, and iterates. Whether that actually happens is still up to the kid and the household norms around the tool — a product can make the iterate-and-evaluate loop easy to reach, but it can't make a kid take it.

No path to the AI playing friend

Common Sense's under-18 line — no AI for emotional support or friendship — applies to every age band we serve, not just the younger ones. In Xyplor, the in-app AI (Nova) is framed to kids as a helper, not a friend or companion; its system prompt forbids role-playing as a friend or family member, and a kid expressing real distress gets redirected to a trusted adult rather than having the AI continue the conversation. This applies the same way whether the kid is 6 or 16, because the guidance doesn't carve out an exception for teenagers on this specific point, and neither do we. (We wrote about the in-app wording change behind this in an earlier post.)

The question worth teaching, regardless of the tool

Common Sense Media's guidance includes a question we think is genuinely useful outside the chatbot context it was written for: "Is this AI helping me learn, or just helping me finish?"

For a chatbot doing homework, the test is fairly direct — did the kid understand the material, or did the bot just hand over an answer. For a maker, the test is a little less obvious because "finishing" looks like real creative output. A kid can type "make me a racing game" and get a playable game in about a minute. That's not the same as a kid who typed that prompt, played the result, noticed the cars were too slow, asked for a fix, noticed the track was boring, asked for a new layout, and kept going until it was actually theirs.

Common Sense also cites a Stanford SCALE Initiative review of more than 800 studies finding that tools which give hints and ask questions outperform tools that hand over finished answers. We're citing that as Common Sense reports it — we haven't independently reviewed the underlying research, and we're not claiming it was conducted on creation tools like ours.

We don't think a product can answer the "learning or just finishing" question for a parent. What a product can do is make the finishing-too-fast version harder to default into and the iterating version easy to reach. Whether a given session was real direction or a one-shot prompt accepted without a second look is still something a parent has to actually look at — which is the whole reason the visible-prompt log exists in the first place.

What we're not claiming

We're not claiming Xyplor teaches critical thinking or media literacy. We're not claiming Common Sense Media has reviewed or endorsed Xyplor — it hasn't, and the guidance in this post is cited as general guidance, not as a rating of any product. We're not claiming visibility and approval gates substitute for a parent actually looking at what their kid made, only that they make looking possible without requiring continuous supervision.

Common Sense Media's guidance is a good baseline for any family evaluating any AI product for a kid, chatbot or otherwise. If you're deciding what "alongside" means for your own household and your own kid's age, their full guidance is worth reading directly rather than taking our summary as the last word on it.

If you want to see what the visibility and approval mechanics described here actually look like in a parent dashboard, we cover the full safety design at xyplor.com/safety, and the age-specific version for younger kids at xyplor.com/ages-6-8.