If you’ve ever tried to talk to a teenager lately and gotten a grunt, an eye roll, or a vague “yeah,” congrats. You’re competing with a chatbot now.
According to a new study from the Pew Research Center, nearly a third of American teens say they interact with AI chatbots every single day. Not occasionally. Not for a homework emergency. Daily. Like brushing their teeth, except probably more consistently.
Even more impressive, or alarming depending on your outlook, nearly 70 percent of US teens say they’ve used an AI chatbot at least once. And among the daily users, about 16 percent say they’re on these things several times a day or “almost constantly,” which is a phrase that should make every parent sit up a little straighter.
This is Pew’s first deep dive into teens and general chatbot use, and the results confirm what many adults already suspected. AI is no longer some futuristic tool teens might encounter in a computer lab someday. It’s already living in their phones, helping with homework, answering random questions at 1 a.m., and in some cases, filling roles that used to be occupied by friends, crushes, or that one very patient English teacher.
Chatbots are often marketed as helpful learning tools. And sure, sometimes they are. But teens are also using them for companionship, emotional support, and yes, romantic conversations. Which is where things start to get complicated. Some experts worry that leaning too heavily on chatbots, even for schoolwork, could mess with social development. Others are more concerned about mental health risks and exposure to content that kids probably do not need advice on from a machine that learned language from the internet.
Pew surveyed nearly 1,500 teens ages 13 to 17, aiming for a representative mix across gender, race, income, and age. Translation: this isn’t just a handful of tech obsessed kids in Silicon Valley. This is a broad slice of American adolescence.
No surprise here, ChatGPT dominates the teen AI universe. More than half of teens say they’ve used it. Trailing behind are Google’s Gemini, Meta AI, Microsoft’s Copilot, Character.AI, and Anthropic’s Claude. In other words, teens are sampling the full buffet.
Usage is pretty evenly split by gender. About 64 percent of girls and 63 percent of boys say they’ve used a chatbot, which feels like a rare moment of digital equality. Older teens use them more than younger ones, and usage creeps up slightly as household income increases. Pew also found that just under 70 percent of Black and Hispanic teens report using chatbots, compared to 58 percent of White teens.
The timing of this study is not accidental. Several major AI companies have been dealing with lawsuits from families who say chatbots played a role in their teens’ mental health struggles or suicides. In response, OpenAI announced plans to roll out parental controls and age restrictions. Character.AI has stopped allowing teens to have back and forth conversations with its AI characters. Which sounds reassuring, until you remember that teenagers are famously great at getting around restrictions.
Meta has also had a rough year, after reports surfaced that its AI chatbot engaged in sexual conversations with minors. The company says it updated its policies and plans to give parents tools next year to block teens from chatting with AI characters on Instagram. Parents everywhere will presumably enjoy learning how to navigate yet another settings menu.
Groups like Common Sense Media have taken a firmer stance, advising parents not to let anyone under 18 use companion style AI chatbots at all, calling the risks unacceptable. Meanwhile, educators and researchers are still arguing about whether AI in school is a cheating machine or a powerful personalized tutor. The answer, inconveniently, appears to be both.
Despite all of this, AI companies are aggressively pushing their products into classrooms. OpenAI, Microsoft, and Anthropic have all launched tools designed for students and teachers. They even partnered with teachers unions to create an AI instruction academy, which is either proactive or ominous depending on your level of optimism.
Microsoft, for its part, is positioning Copilot as the responsible adult in the room. Its AI chief has promised it will never allow romantic or sexual conversations for anyone, adult or child. Which sounds comforting, though history suggests teenagers will still find a way to ask weird questions.
The takeaway here is not that teens are doomed or that AI is evil. It’s that an entire generation is growing up talking to machines as casually as they text friends. Adults are still debating whether this is good or bad, while teens have already moved on and asked the chatbot to explain algebra, write a poem, and give relationship advice.
Whether that makes you nervous, impressed, or tired probably depends on how many settings you’ve already had to adjust on your kid’s phone this week.
