72% of Teens Are Already Talking to AI Companions. Is Your Kid One of Them?

Your child is using AI not just to get homework answers, but also to talk, vent, ask for relationship advice, seek mental health support, and in some cases, find companionship.

Currently, approximately 72% of teenagers use an AI virtual assistant. This data comes from the Center for Democracy and Technology’s 2025 report on schools and AI risk. Pew Research Center corroborates it: 64% of teens use AI chatbots regularly, and 3 in 10 interact with one every single day.

A significant number of middle schoolers are interacting daily with something that is always available, always willing to listen, and never gets tired of hearing them out, and all of this is happening during the most emotionally turbulent, identity-forming, and socially complex years of childhood. The scariest part is that most parents have no idea this is already happening in their own homes.

We are talking about AI virtual assistants like Character.AI, Snapchat’s My AI, and Replika, which are designed to simulate a relationship. They validate instantly, ask follow-up questions, and make kids feel heard without any of the social risks that come with real human connection. For a middle schooler who does not want to burden their parents or risk being judged by a friend, that is a powerful pull.

Research shows that teens are using AI companions for mental health support (42%), friendship and companionship (42%), escaping real life (42%), and romantic relationships (19%). These are patterns repeating themselves in ordinary households, on ordinary evenings, without most parents knowing.

Today, only 1 in 10 teachers have been trained on how to respond when AI use affects student wellbeing. This tells us that the adults closest to these kids, both at school and at home, are mostly unprepared for what is already happening, and the tools children are turning to were not designed with their developmental needs in mind.

This is a home problem that also shows up at school, not the other way around. The hours that matter most, after school, late in the evening, those quiet moments when a child reaches for their phone instead of talking to someone, those hours belong to parents.

A 2026 Rithm Project study of nearly 2,400 young people found that what protects kids from high-risk AI use is not a technical filter or a school policy. It is an authentic, safe human connection. Their conclusion: non-judgmental intergenerational conversation may be the most important intervention available.

This is a parent intervention. It means you and me. Not a monitoring app. Not a new device rule. A conversation. And if you do not know how to start it, begin with one question:

“Have you ever talked to an AI chatbot for advice, like ChatGPT? What did you use it for? What kind of response did you get and how did it feel?”

Curious, not interrogating. If they say no, ask if any of their friends have. You will often get more from that answer. The goal is not to catch them doing something wrong, but to become the kind of parent they come to instead of a chatbot.

Awareness is the first step, and most parents I talk to want to know what to actually do with that awareness. That is exactly what PARENT-IN-THE-LOOP™, AI education and training for parents, was built for, with the tools to know, the conversations to have, and a practical path forward for parents navigating this in real time.

If you want to go deeper on this topic, I am building a live virtual workshop specifically for middle school parents. Leave a comment below or reach out directly. I want to make sure what I build is what you actually need.

From awareness to agency to advocacy.

Julie | LIKEAMOTHER.AI™