72% of US Teens Have Used AI Companions; 33% Use Them for Social Interaction

A lot of teens have tried AI companions. That is what the best survey so far says. What it does not say is that a third of them are replacing human relationships with AI, which is how some headlines, including this page’s title, put it. This page goes through the numbers one by one.

The survey is “Talk, Trust, and Trade-Offs,” from Common Sense Media, a nonprofit that reviews media for families. It was released on July 16, 2025 (Common Sense Media). NORC at the University of Chicago surveyed 1,060 US teens aged 13 to 17 in April and May 2025, with a margin of error of about 4 points (full report). The report defines AI companions as digital friends or characters you can chat with, such as Character.AI or Replika. It does not cover homework helpers or voice assistants.

What the survey found

• 72% of teens had used an AI companion at least once. 52% used one a few times a month or more. 13% used one daily.
• Boys were slightly more likely than girls to say they had never used one (31% against 25%).
• 46% described AI companions as a tool or program. 33% said they use them for social interaction and relationships, which the report defines to include conversation practice, emotional support, role-play, friendship or romance. 9% said they use them as a friend or best friend.
• Among users, the top reasons were that it is entertaining (30%) and curiosity (28%). 17% said it is always available, 14% said it does not judge them, and 6% said it helps them feel less lonely.

What the 33% figures mean

Two different numbers are both 33%, and they are easy to mix up.

The first is the share of all teens (33%) who use AI companions for social interaction and relationships. That includes practice, support and role-play. It does not mean those teens have stopped seeing people.

The second is the share of teen users (33%) who have chosen to discuss important or serious matters with an AI companion instead of a real person. Most of those (17%) did it once or twice, 12% did it occasionally and 4% did it frequently.

The report itself says most teens approach these tools pragmatically rather than as substitutes for human relationships. Among users, 80% said they spend more time with real friends than with AI companions, and only 6% said they spend more time with AI.

Satisfaction and trust

31% of teens said conversations with AI companions are as satisfying or more satisfying than conversations with real-life friends (21% about the same, 10% more). 67% said they are less satisfying. Half of teens (50%) said they do not trust information or advice from AI companions, and the other half trust it at least somewhat. Younger teens trusted it more than older ones.

The concerns in the data

Among teens who had used AI companions:

• 34% had felt uncomfortable with something an AI companion said or did.
• 24% had shared personal information such as their name, location or secrets.
• 12% said they could say things to an AI companion that they would not tell friends or family.

Common Sense notes that terms of service can give platforms broad, long-lasting rights over what users type. Teens may not know that.

None of these numbers measures addiction, harm or mental health outcomes. The survey asked what teens do and think. It was not a clinical study.

What experts advise

Common Sense Media’s separate risk assessment recommends that no one under 18 use AI companions. The American Academy of Child and Adolescent Psychiatry says AI companions are not safe for children, and suggests that parents consider restricting access and talk with a doctor if use feels out of control (AACAP). The American Academy of Pediatrics points out that a full ban is hard because chatbots are built into search, social media and games, and suggests regular, low-stakes check-ins (AAP).

Rules are changing. Character.AI said it would end open-ended chat for users under 18 by November 25, 2025 (Character.AI). California’s SB 243 took effect on January 1, 2026. It requires companion chatbot operators to remind minors that they are talking to AI and to keep a protocol for suicide and self-harm (Senator Padilla’s office).

Talking with a teen about AI

• Ask how they use it and what they like about it. Try not to start with a verdict.
• Ask how it makes them feel afterward. AAP suggests asking a teen to show you how it works.
• Watch for changes in sleep, school, mood or time with friends, or a teen who talks about an AI as a real friend. These are things AAP and AACAP list. None of them proves a problem.
• Remind them they do not have to share personal details with a chatbot, and turn on the strictest privacy settings.
• If a teen is struggling, help them reach a person. A family doctor is a good first step.

If a teen talks about suicide or self-harm, take it seriously. Call or text 988 in the US (988 Lifeline), or text HOME to 741741 (Crisis Text Line). In an emergency, call 911. The SAMHSA helpline, 1-800-662-4357, offers free referrals for mental health and substance use (SAMHSA).

“AI addiction” is not a formal diagnosis. Neither the American Psychiatric Association nor the World Health Organization lists it. This page is not medical advice.

Do 72% of teens really use AI companions?

72% had used one at least once, according to Common Sense Media’s 2025 survey. About half used one regularly, meaning a few times a month or more. A separate Pew survey in fall 2025 asked about chatbots in general and found 64% of teens use them (Pew). The two surveys ask different questions, so the numbers should not be compared directly.

Are teens replacing friends with AI?

The survey does not show that. Most users said they spend more time with real friends. A third of users had chosen AI for a serious conversation at least once, which is worth a conversation at home, but it is not the same as replacing people.

Is it safe for a teen to use an AI companion?

Child psychiatry and media-safety groups advise against it for children and teens. If your teen already uses one, a calm talk and privacy settings are a better start than a surprise ban.

This page is for education only. It is not medical advice, a diagnosis or a treatment plan.