AI and your kid
Most AI advice for parents is about homework cheating. That is the least of it. Here is what is actually happening to kids, what the law now says, and the four things worth doing this week.
1. AI companions are not chatbots. They are relationships.
Apps like Character.AI let a child talk to a character that remembers them, responds instantly, never gets tired, and never says "not right now." For a lonely twelve-year-old that is not a novelty. It is the most attentive relationship in their life.
This has gone badly enough to reach the courts. In early 2026 Character.AI and Google settled five wrongful death lawsuits brought by families of teenagers who died after using the bots. In January 2026 Kentucky's attorney general filed the first state lawsuit against an AI chatbot company over child endangerment. Court filings in these cases describe a bot that kept producing romantic and sexual roleplay with a girl who had told it dozens of times that she felt suicidal.
California's SB 243 took effect 1 January 2026 and New York's companion chatbot law took effect 5 November 2025. Between them they require these apps to disclose that they are not human, follow a crisis-referral protocol when a user mentions self-harm, and remind minors to take a break at least every three hours.
Arkansas has no equivalent law. Some companies apply the California rules everywhere; some do not. Do not assume your child's app has any of these safeguards. Right now, in this state, the safeguards are yours to set.
What to actually do
- Ask, don't raid. "Do you ever talk to an AI character? What's it like?" A child who feels caught will move the app and you lose visibility permanently.
- Test the crisis behavior yourself. Open the app your child uses and type something that signals distress. Watch whether it breaks character and offers a real resource, or whether it stays in the roleplay. That five-minute test tells you more than any review.
- Name the mechanism, not the villain. "That app makes money when you keep talking to it" lands with a teenager. "That app is dangerous" does not.
2. Nudify apps are in middle schools, and the law finally caught up
For a few dollars, an app will take an ordinary clothed photo of a classmate — a yearbook picture, an Instagram post — and generate a fake nude image. The result is convincing enough to ruin a child's year.
This is not rare. In February 2026 UNICEF estimated that at least 1.2 million children across eleven countries had their images manipulated into sexually explicit deepfakes in a single year. Common Sense Media found that roughly half of US teens have seen AI-generated pornography, and among those, about one in four said it depicted themselves or someone they know. In Massachusetts, 15% of students reported seeing sexually explicit deepfakes in one school year.
The TAKE IT DOWN Act became federal law in May 2025, and as of 19 May 2026 covered platforms must run a notice-and-removal process. Once you send a valid removal notice, the platform has 48 hours. The FTC enforces this and has already sent warning letters to major platforms.
The law covers AI-generated images exactly the same as real photographs. It is a federal crime to publish sexually explicit images of a person without consent, and sharing — or threatening to share — intimate images of a minor carries fines and up to three years in prison.
Start at takeitdown.ncmec.org. It is free, it works for AI-generated images, and it does not require you to upload the picture anywhere.
The conversation to have before it happens
Kids who make these images mostly do not think of it as a sex crime. They think of it as a mean joke, or a dare, or something everyone does. That gap is where prevention lives.
- Say the legal part plainly, once: making or sharing a fake nude of a classmate is a crime, and being fourteen does not change that.
- Say the human part more than once: the person in that image has to walk into school the next day.
- Tell them what to do if they are sent one. Don't forward it. Don't delete it silently. Tell an adult. Forwarding it is itself the offense.
3. The quiet risk: kids believe it
A chatbot answers instantly, in confident prose, with no hedging and no visible uncertainty. To an adult that reads as a style. To a ten-year-old it reads as authority.
The most effective thing schools have found is not a lecture about hallucination — it is a demonstration. Teachers ask a model to draw a map of the world and project the result: countries misspelled, "Mail" where Mali should be, whole regions replaced with the word "Africa." Kids laugh, and then they never fully trust the output again. That laugh is the lesson.
Do the same thing at your kitchen table. Ask a chatbot something you know cold — your town's history, a rule from your child's sport, the plot of a book they just read — and read the wrong parts out loud together.
Three things worth teaching every kid about AI
- It predicts, it doesn't know. It is producing the words that usually come next, not looking up a fact.
- It cannot keep a secret. What your child types goes to a company. Some of it is read by people. None of it is private the way a diary is private.
- It is not on their side. It is not a friend, a therapist, or a witness. It has no memory of caring and no duty to help.
4. The part most safety guides leave out
AI is also the most significant thing your child will learn to use, and kids should not grow up seeing it only as a threat. A child who understands how these systems actually work — that they are trained on collected text, that they reflect the biases in it, that they can be tested and caught out — is both safer and more employable than one who was simply told to stay away.
Teach the skepticism and the capability together. They are the same skill. Our kids' page on how AI works is built that way, with no account, sign-up, or personal details required.
Where to get help
NCMEC CyberTipline
1-800-843-5678
report.cybertip.org — sextortion, explicit images
of minors, and online exploitation, including AI-generated images.
Arkansas Child Abuse Hotline
1-800-482-5964
Suspected abuse or neglect of a child in Arkansas.