Seven in Ten Teens Have Tried an AI Companion. Many Parents Are Not Sure Their Teen Uses Chatbots at All.

When researchers at Drexel University's ETHOS Lab asked teenagers to describe how they came to rely on AI chatbots, the stories tended to begin in the same unremarkable place. "Many teens described starting with something that felt helpful or harmless, but over time it became something they struggled to step away from, even when they wanted to," doctoral student Matt Namvarpour said in an April announcement of the lab's findings.

In a national survey of 1,060 teens conducted in the spring of 2025, Common Sense Media found that 72 percent had used an AI companion at least once and 52 percent used one at least a few times a month. A third said they had chosen an AI companion over a person for a serious conversation, according to its report Talk, Trust, and Trade-Offs.

For parents, the evidence points toward a calm, curious conversation rather than a ban or a lecture. Laws in New York and California and new company safeguards now set a floor of protection. The research suggests that what happens at the kitchen table still shapes how a teenager uses these tools more than any setting in an app.

What the Surveys Show About Teens and AI Companions

The Common Sense survey, fielded by NORC at the University of Chicago, found that 13 percent of teens used an AI companion daily and 21 percent a few times a week. Thirty-one percent said conversations with an AI companion were as satisfying as or more satisfying than conversations with real friends. Another 34 percent said a companion had said something that made them uncomfortable, and 24 percent had shared personal information with one.

The same report offers some reassurance, since 80 percent of teens said they spend more time with real friends than with AI companions, and half said they do not trust the advice companions give. Younger teens were more trusting, with 27 percent of 13- and 14-year-olds saying they trusted companion advice quite a bit or completely, compared with 20 percent of 15- to 17-year-olds.

The Pew Research Center, which asks about chatbots more broadly, found in a December 2025 report that 64 percent of teens use AI chatbots and roughly three in ten use them daily. In a February 2026 follow-up, 12 percent of teens said they use chatbots to get emotional support or advice.

The Gap Between Teens and Parents

Parents often do not see the whole picture. In Pew's parallel survey of parents, 51 percent said their teen uses chatbots and about three in ten were unsure, a figure 13 points below what teens reported about themselves. Only 18 percent of parents said they were comfortable with their teen seeking emotional support or advice from one.

A newer Common Sense report released in June, covering children and teens from 9 to 17, found that "nearly half of kids have never had a conversation with their parents about AI safety," according to the organization's research summary. The group's position is blunt, stating that AI companionship is not safe for anyone under 18.

Why Psychologists Are Paying Attention

"People don't realize that relational learning happens during the teenage years and that these moments of social connection are little building blocks that become bigger things that will benefit you throughout life," said Thao Ha, an Arizona State University psychologist, in a June announcement of her paper in The Lancet Child and Adolescent Health. "You really need those building blocks so you actually learn the skills that you need to thrive in your relationships."

Ha also pointed to the pace of change. "The technologies are developing super fast, faster than we can keep up with as scientists, faster than governance and policy can keep up with," she said.

The American Psychological Association issued a health advisory on artificial intelligence and adolescent well-being in June 2025. Its recommendations, as published by ScienceDaily, begin with "ensuring there are healthy boundaries with simulated human relationships," noting that adolescents "are less likely than adults to question the accuracy and intent of information offered by a bot." Mitch Prinstein, the association's chief of psychology, framed the advisory as practical. "Many of these changes can be made immediately, by parents, educators and adolescents themselves," he said.

The companion question sits alongside broader teen mental health pressures that schools and families are already navigating, which TheStandardNY explored in its guide to responding to a teen's mood before a crisis.

What the New Laws and Company Rules Require

Since November 5, 2025, AI companion safeguards in state law require these products to detect expressions of suicidal thinking or self-harm and "promptly direct users to crisis service providers." They must also tell users they are not speaking with a human at the start of a session and "at periodic intervals of every three hours of continued companion use," according to Governor Kathy Hochul's letter to AI companion companies. The attorney general enforces the law.

A second New York bill, S9051B, would bar companion features aimed at minors that simulate a romantic relationship, claim human emotions or encourage secrecy. It passed the Senate 60 to 0 on June 4 and the Assembly the next day. The Senate lists it as delivered to the governor, with a January 1, 2027 effective date if it becomes law.

California's SB 243, which took effect January 1, 2026, requires companion platforms to remind minors that they are talking with AI, block sexually explicit content for minors and maintain crisis protocols. On September 10, Governor Gavin Newsom signed a broader measure, SB 1119, that adds age assurance, default parental controls and independent child safety audits, according to the governor's office. At the federal level, the Federal Trade Commission opened an inquiry in September 2025 into seven companies, including Alphabet, Character Technologies, Meta, OpenAI, Snap and xAI, and has not released findings.

Changes Inside the Apps

Character.AI announced in October 2025 that it would end open-ended chat for users under 18, and the removal began in stages on November 24. OpenAI introduced parental controls for ChatGPT in September 2025 that let linked parents set quiet hours, turn off memory and voice mode, and receive alerts if there are signs a teen may be considering self-harm. Parents cannot read the conversations, and on August 18 the company launched ChatGPT for Teens, which automatically enrolls users it estimates are under 18 and blocks romantic roleplay.

These protections vary by platform and depend on accurate age detection, which is why the settings on a teen's phone matter as much as the policy on paper. Our look at what Meta's settlement changes on your teen's phone walks through how those defaults work in practice.

How to Start the Conversation at Home

The Common Sense report recommends opening the subject by asking, without judgment, which platforms a teen uses and how those exchanges feel compared with human friendships. The American Academy of Pediatrics suggests similarly low-stakes prompts such as asking how a teen is using AI.

Teens often respond well to an honest explanation of how these products are designed. Common Sense advises parents to "explain that AI companions are designed to be engaging through constant validation and agreement," so a teenager can tell a chatbot that always agrees from a friend who sometimes pushes back.

Parents should also know the warning signs of unhealthy AI use, which the report lists as social withdrawal, declining grades and a preference for AI companions over human interaction. Pediatricians add changes in sleep or regular activities. A family agreement about where devices charge overnight, reached during a calm moment such as the twenty-minute family meeting we have written about, tends to hold better than rules announced after a conflict.

If a teen ever mentions thoughts of self-harm, in a chat log or in person, the 988 Suicide and Crisis Lifeline takes calls and texts at any hour. A pediatrician or school counselor can help a family decide on next steps.

What Families Can Expect Next

More rules are coming, with New York's S9051B awaiting action and California's new law adding audits and parental controls that companies will have to build. Congress has also moved on children's online safety, and TheStandardNY has reported on how New York's 2024 laws anticipated the House bill.

Common Sense found that 80 percent of teens still spend more time with real friends than with AI companions. In the same survey, 33 percent had already brought a serious conversation to a chatbot instead of a person.