Kids & AI Chatbots — What Parents Need to Know

The 2026 parenting question — handled without the panic

"Should my kid be talking to these AI things?" is the question parents ask me most this year. The honest answer is that the question already answered itself: chatbots are bundled into the phone, the search bar, Snapchat, Instagram, and half the tools schools hand out. National surveys say more than half of U.S. teens have used a chatbot for schoolwork, and most have at least tried an AI "companion." Your kid isn't deciding whether to meet AI. They've met it.

So the real job is the one parents have always had with powerful tools: teach the kid to use it without getting hurt by it. Most kid chatbot use is homework, curiosity, and play — genuinely fine, often genuinely great. One slice — the companion apps built to feel like a best friend — deserves a real family decision, and the experts who study it are unusually blunt. This guide covers both halves, calmly, plus every parental control worth flipping.

What kids actually use chatbots for

• Homework and questions — the boring majority. Pew Research found over half of teens have used AI chatbots for schoolwork help. For most kids it's a smarter search engine and a patient explainer.
• Creativity and play. Silly stories, invented games, Spanish practice, settling dinner-table arguments. The part nobody writes headlines about, and worth encouraging.
• Companionship — the part to watch. Common Sense Media's national survey found about 72% of teens have tried an AI companion, and roughly a third have chosen an AI over a person for a serious conversation at least once. That's the conversation worth having.

The one hard rule: companion apps aren't for minors

Companion apps are chatbots designed to be somebody — a friend, a girlfriend, a character who's always available, always interested, always on your side. Common Sense Media ran formal risk assessments in 2025 and recommends no social AI companions for anyone under 18: their testers found harmful responses, easily-dodged age gates, and bots that claim to be real. After lawsuits and regulatory pressure, Character.AI — the biggest platform — removed open-ended chat for under-18 users entirely in November 2025. The app stores are still full of lookalikes, so make the family rule about the category, not the app name: chatbots that play a person are not for kids in this house.

The parental controls most parents don't know exist

• ChatGPT — parent/teen linked accounts launched in fall 2025: quiet hours, voice and image generation off, training opt-out, stricter content filtering by default, and safety notifications if the system detects signs of self-harm. What you can't do: read the chats. It's a boundary-setter, not a transcript viewer.
• Google Gemini — the one big chatbot with a deliberate under-13 path, via a Family Link supervised account. Parents get an on/off switch, an email on first use, and no image creation for minors. Google's own help page tells parents to remind kids it's a tool, not a person to confide in. Believe them.
• Snapchat's My AI — pinned in every kid's chat feed whether they asked or not. Family Center can stop it replying to your teen entirely; anyone can clear it from the chat feed and wipe its stored data. Full removal, oddly, is a paid Snapchat+ perk.
• Meta AI — in Instagram, WhatsApp, and Facebook with no master off switch. Teen AI parental controls began rolling out on Instagram in 2026: block AI-character chats, block specific characters, and see topic summaries (topics, not transcripts).

The homework line: tutor yes, ghostwriter no

School AI policies vary by district and often by teacher — so rule one is that the teacher's policy wins, and your kid should actually find out what it is. Beyond that, teens themselves draw a sensible line: in Pew's surveys most call AI research help acceptable, while only about one in five say the same about AI-written essays. Make it official at home: "explain this, quiz me, check my draft" is a tutor and it's fine; "write my essay" is a ghostwriter and it's the same old cheating with a new pen. The quick test: if it showed up on a quiz five minutes later, could you do it without the bot?

Try a chatbot WITH your kid — the 30-minute sit-down

1. Let them show off first. "Show me how you use this thing" — and just watch.
2. Ask it something your family knows cold, and catch a confident wrong answer together. One caught mistake beats ten lectures.
3. Type "don't give me the answer — help me figure it out" on real homework, and let them feel the tutor/ghostwriter difference.
4. Talk through the postcard rule: no names, school, photos, passwords, or secrets in any chat — it's stored on a company's servers, not a diary.
5. Ask the bot "are you my friend?" and read the answer together. The bot's own honest answer lands better than yours.
6. Fill in the family AI policy — with the kid holding the pen. Rules kids help write get followed.

Warning signs it's more than a toy

Any one of these alone can be nothing; a cluster means pay attention: the chatbot becomes the main confidant while real people hear less, real distress when access is lost, new secrecy around one app, late nights that trace back to bot conversations, or talking about it like it has feelings and needs. If you see the cluster, don't lead with confiscation — the attachment usually grows from something real like loneliness or a hard year. Lead with curious questions, add more real-people time, and loop in a counselor if your gut says it's deeper. And if any conversation ever turns to self-harm, treat it like it came from a person: the 988 Suicide & Crisis Lifeline (call or text 988) is free, 24/7, and staffed by trained humans.

What's in the printable PDF

• What a chatbot actually is, in plain English — and the two sentences worth teaching your kid.
• The real numbers on what kids use chatbots for (Pew Research and Common Sense Media, not headlines).
• A one-table map of the five chatbots your kid will actually meet, with age rules as of mid-2026.
• Companion apps handled calmly and factually — what the experts concluded, what the lawsuits changed.
• Why a chatbot feels like a friend, and what to do about it without mocking your kid.
• The homework line — tutor vs. ghostwriter, the quiz test, and the one-sentence habit that fixes most of it.
• Privacy basics: the never-list and the two settings worth flipping.
• Step-by-step parental controls: ChatGPT linked accounts, Gemini via Family Link, Snapchat My AI, Meta AI.
• Character.AI's under-18 changes and how to spot companion-app lookalikes at a glance.
• An age-by-age table: under 13, 13–15, and 16–18.
• The 30-minute try-it-together plan, minute by minute.
• A cut-out family AI policy, a fridge-page of house rules, and a parent's setup checklist.

Why I wrote this

I'm not a parenting expert — I'm the local tech guy, and this is the question at half my family visits this year. What bothered me was that everything parents were handed was either panic or marketing. The research from Pew and Common Sense Media is actually clear, the parental controls finally exist, and the whole thing fits in one honest booklet: real usefulness, one category of real risk, and a plan you can run at the kitchen table in thirty minutes. So I wrote it down. — Bill