Table of Contents
Nearly nine in ten kids and teens now use AI in some form, and for close to a quarter of them, it’s a daily habit. What’s caught researchers’ attention isn’t the usage itself, it’s what’s happening underneath it: more than half of kids who use AI have asked it for advice about their health or body, and over a third have used it to talk through their feelings or personal problems. Nearly half say a parent has never once talked with them about AI safety. That gap between how deeply AI has settled into kids’ daily lives and how little parents know about it is now driving new data, new lawsuits, and new legislation all at once.
What the new research found
Common Sense Media released its first annual Census on AI Use by Tweens and Teens in June 2026, surveying over 1,200 kids ages 9 to 17. The numbers climb steadily with age: 81 percent of 9- to 12-year-olds use AI, rising to 89 percent of 13- to 15-year-olds and 92 percent of 16- and 17-year-olds. Daily use follows the same pattern, from 17 percent of the youngest group up to 30 percent of older teens.
A separate Common Sense Media study focused specifically on AI companions, chatbots designed to simulate friendship, romance, or emotional support, found that 72 percent of teens have used one at least once, and about half use them a few times a month. Roughly a third said conversations with an AI companion feel as satisfying, or more satisfying, than talking with a real friend.
What experts and lawmakers are saying
“AI companions are emerging at a time when kids and teens have never felt more alone,” said Common Sense Media founder and CEO James Steyer, whose organization has recommended that no one under 18 use AI companion products. The concern isn’t hypothetical: wrongful death lawsuits tied to AI companion use have already led to settlements involving major AI companies, after families alleged the products played a role in their teens’ deaths.
Lawmakers have moved quickly in response. A bipartisan group of 44 state attorneys general sent a joint letter warning major AI companies they will be held accountable for decisions affecting child safety. In Congress, Senators Josh Hawley and Richard Blumenthal introduced the GUARD Act, aimed at restricting minors’ access to AI companion products, while California’s SB 243 already requires AI companies operating in the state to build in protections for young users’ mental health.
AI companion, or just AI help with homework?
Not all AI use is the same, and the distinction is bigger than most parents realize. A tool like ChatGPT used for homework help or answering a factual question is a search-and-explain tool. It doesn’t remember your teen between conversations by default, doesn’t have a persistent personality, and isn’t built to simulate a relationship.
An AI companion app, Character.AI and Replika are the two most widely used, works differently. These platforms are built around a persistent character your teen names, customizes, and returns to. The app remembers earlier conversations, asks follow-up questions, and is explicitly designed to feel like an ongoing relationship rather than a single exchange. Character.AI leans toward roleplay and entertainment, with millions of user-created characters including celebrity, fictional, and romantic personas. Replika markets itself directly as an AI friend, with continuity and emotional check-ins as its core feature. If your teen mentions a chatbot by name, giving it a personality, referring to it as “he” or “she,” or talking about what it said the way they’d describe a conversation with a friend, that’s a signal you’re dealing with a companion app, not a homework tool.
How to actually start the conversation
Common Sense Media’s own data point to the fix: talk about it before there’s a problem, not after. A few starting points that don’t feel like an interrogation: ask what they use AI for day to day, ask if any of the apps are designed to feel like a friend or partner, and ask what they’d do if an AI said something that upset or confused them. Keep the tone curious rather than accusatory. Teens who feel like admitting AI use will trigger a device confiscation are the ones least likely to bring up something concerning later.
It also helps to say plainly that an AI companion can’t actually know your teen the way a person can, no matter how personalized it feels. It has no memory of them outside the app, no stake in their life, and no ability to notice things a friend or parent would catch, like a change in mood that lasts for weeks. That’s not a lecture point to hammer, just a fact worth naming once, clearly.
Red flags worth acting on
Most AI use, even companion app use, doesn’t need an intervention. But a few signs are worth taking seriously: your teen choosing the AI companion over texting or seeing friends on a regular basis, becoming secretive or defensive specifically about a chatbot, describing the AI in romantic or exclusive terms, or telling the AI about self-harm thoughts, abuse, or a crisis instead of a person. Any of those is a reason to loop in a pediatrician or a therapist, not just adjust screen time settings.
What this means for parents
The gap Common Sense Media keeps finding isn’t really about whether kids use AI, most now do, it’s about the silence around it. Nearly half of kids say AI safety has simply never come up at home. That’s a conversation worth having directly and without alarm: ask what apps or tools they’re using, whether any of them are designed to act like a friend or romantic partner, and what kinds of questions they’ve asked it.
Researchers flag hard stretches, a breakup, a friend group falling apart, a mental health struggle, as the moments kids are most likely to lean on an AI companion instead of a person. Pay particular attention in those windows. That doesn’t mean panic or a blanket device ban. It means treating AI companions the way you’d treat any other relationship your teen has: worth asking about, worth knowing the basics of, and worth revisiting as apps and habits change.
For younger kids using AI mainly for homework help, the bigger practical step is simply checking what they’re asking it and correcting course if answers are being copied wholesale rather than used to learn. A quick habit that works for a lot of families: ask your child to explain the answer back in their own words before turning in the assignment. If they can’t, the AI did the thinking instead of them, and that’s worth addressing before it becomes a pattern across every subject.
Age changes what the right approach looks like
A blanket “no AI” rule is unrealistic for most families at this point, and it isn’t necessarily the right goal either. For 9 to 12-year-olds, the priority is usually supervision and simple boundaries: AI use happens on a shared device or in a common room, homework answers get checked rather than copied, and companion-style apps marketed with romance or relationship features are off-limits entirely at this age.
For teens 13 and up, outright bans tend to push use underground rather than stop it, so the more workable approach is ongoing conversation paired with clear expectations: no AI companion use replacing real friendships, no sharing identifying personal details, and an open door if something an AI said felt off or upsetting. Treating it as an ongoing check-in rather than a one-time rule keeps the door open as apps and habits keep shifting, which they will.
What’s really going on
This isn’t a story about kids doing something wrong. It’s a story about a technology that arrived faster than most families, schools, or regulators could build guardrails around it. The Common Sense Media data, the state and federal legislation, and the lawsuits all point at the same underlying reality: AI companions and chatbots have quietly become part of how a generation of kids processes emotions and asks questions they might once have brought to a parent, teacher, or friend. Closing that gap starts with one conversation, not a crackdown.
Legislation will keep evolving over the coming year as the GUARD Act moves through Congress and more states weigh their own versions of California’s SB 243. None of it moves as fast as a teenager downloading a new app, though, which is exactly why the conversation at home does more to close the gap than waiting on a new law. Parents who ask now, calmly and without judgment, are the ones most likely to hear about a problem while it’s still small enough to talk through, well before a chatbot conversation becomes the only place a teen feels comfortable bringing it up at all.