The AI Conversation Gap Between You and Your Teenager
You probably think you have a decent read on whether your teenager uses Claude or ChatGPT. New data from Pew Research says most parents are wrong, and not by a little.
The Gap Parents Don’t Know They Have
Pew surveyed more than 1,400 US teens and their parents this past fall. 64% of teens say they've used AI chatbots. Only 51% of parents think their teen does. Almost three in ten parents said they honestly don't know (Pew Research Center, 2026). That's not a rounding error. That's close to a third of parents with no read at all on something their kid is doing regularly.
The gap isn't the technology. The gap is that a third of parents have no idea what their kid actually does with it.
What Teens Actually Think, And It’s More Complicated Than The Headlines
Teens are not uniformly for or against this the way coverage keeps framing it. 36% think AI will be personally good for them over the next twenty years, only 15% think it'll be personally bad. Ask them about society instead of themselves, and they get noticeably more skeptical, the positive number drops and the negative number climbs. 59% think AI-assisted cheating is a regular occurrence at their own school right now (Pew Research Center, 2026). That's not a teen who's naive about this technology. That's a teen holding a more complicated, more accurate view of it than a lot of adults do.
What Unhealthy Use Actually Looks Like
Researchers have started describing unhealthy AI use in teens using the same framework used for behavioral dependence generally, the components model of addiction (Griffiths, 2005). Six markers:
conflict (wanting to cut back and not being able to),
salience (the AI relationship starting to crowd out real ones),
withdrawal (real distress when it's unavailable),
tolerance (needing more of it for the same relief),
relapse (quitting and going back),
and mood modification (reaching for it specifically during stress or loneliness).
Most teen AI use is ordinary. The six markers describe something else entirely, and most kids aren't in that picture.
The American Psychological Association has issued a health advisory on artificial intelligence and adolescent wellbeing (APA, 2025). Common Sense Media, working with Stanford's Brainstorm Lab for Mental Health Innovation, tested the major AI companion platforms directly and found they are not currently safe tools for teens seeking mental health support, and recommends against AI companions for anyone under 18 (Common Sense Media, 2025).
What This Doesn’t Mean
Your teenager using Claude or ChatGPT for a history paper is not the picture above. Most teen AI use, per Pew's own numbers, is searching for information and getting help with schoolwork: ordinary, low-stakes, nowhere near the six markers.
The Reframe Most Parents Are Missing
Here's the part that actually changes how I'd have you think about this. The instinct right now, understandably, is fear: kids won't know how to think anymore, AI will do it for them, the muscle will just atrophy. I'd push back on that. The people who hand AI the entire job, let it think, write, or feel for them completely, are the ones who lose something real. Not because the tool is dangerous. Because they never built the muscle to begin with. The people who use it to sharpen what they're already doing, faster drafts, cleaner thinking, more precise work, tend to come out ahead. Not despite the AI. Because of how deliberately they used it.
The kids who let AI think for them completely are the ones who lose. The ones who use it to think more precisely are the ones who win.
This isn't new territory for us as parents, either, if you zoom out. Education has been moving away from one-size-fits-all for years, you can see it in the range of school models now available, in curriculum choices, in the entire conversation around personalized learning. We already accept that kids don't all learn the same way. AI is just the next place that same lesson shows up, and the parents who treat it as a tool to be shaped, not a threat to be feared or a shortcut to be indulged, are the ones setting their kids up to use it well.
What To Actually Do
You may not currently know which category your kid is in, because most parents don't. The way to find out isn't surveillance, it's a specific kind of conversation: curious instead of accusatory, before there's a reason to be worried, not after. Ask what they use it for. Ask what they'd tell it that they wouldn't tell you, and sit with that answer instead of reacting to it. Talk about where the things they type actually go, because it isn't nowhere. Help them notice that an AI is built to agree with them in a way a real friend, with actual opinions, never will.
This is the same move you'd make with any new technology, screens, social media, a car. Not "is it good or bad," but "how do we build a relationship to it that leaves room for both of us to say something's off."
What Happens Next
Reading this gives you the research. It doesn't give you the read on your specific kid, what their actual relationship to AI looks like right now, whether the conversation needs to happen this week or can wait, or what "curious instead of accusatory" sounds like for the particular teenager in your house who shuts down at the first hint of a lecture.
That's the part that doesn't come from an article, no matter how thorough. It comes from someone mapping your kid's specific patterns with you, not handing you a generic script. More posts on this won't get you there. A discovery call is where we actually build that map together, thirty minutes, and you leave knowing exactly what to watch for and what to say first.
By Lauren Greeno
Child & Adolescent Development Specialist & Parenting Coach | Founder, The Parenting Collaborative
Lauren specializes in helping parents understand invisible dynamics shaping their children’s development and redesigning family systems before patterns calcify into adult identity. With expertise in child development, family systems theory, and trauma-informed parenting, she works with families navigating sibling dynamics, only child considerations, neurodivergence, emotional regulation, and breaking generational patterns.
Work with Lauren: Book a discovery call | Learn more| Instagram | TikTok
References:
American Psychological Association. (2025). Artificial intelligence and adolescent well-being: An APA health advisory. APA.
Common Sense Media. (2025). AI risk assessment: Social AI companions. Common Sense Media, in partnership with Brainstorm: The Stanford Lab for Mental Health Innovation.
Griffiths, M. D. (2005). A "components" model of addiction within a biopsychosocial framework. Journal of Substance Use, 10(4), 191-197.
Pew Research Center. (2026, February 24). How teens use and view AI. Pew Research Center.
Pew Research Center. (2026, February 24). What parents say about their teen's AI use. Pew Research Center.
