72% of US Teens Have Used AI Companions: What the Common Sense Media Survey Found

If you or someone you know is in crisis, call or text 988 (the 988 Suicide and Crisis Lifeline in the US) or text HOME to 741741 to reach Crisis Text Line. Both are free and open at all hours. This page is general information, not medical advice.

Published: August 14, 2025

What the Common Sense Media Survey Found

Common Sense Media, a US nonprofit, published a survey report on July 16, 2025 called Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions. The headline number is that 72% of US teens aged 13 to 17 had used an AI companion at least once. This post reports what that survey says and what it does not.

The data were collected by NORC at the University of Chicago, with funding from Common Sense Media. NORC surveyed 1,060 teens between April 30 and May 14, 2025. The report gives a margin of sampling error of plus or minus 4.2 percentage points overall, and says it may be higher for subgroups. The report’s methodology section has the full details (PDF).

A few numbers from the report, all as the report states them:

• 72% of teens had used AI companions at least once.
• 52% were regular users, meaning a few times a month or more.
• 13% used them daily (8% several times a day, 5% once a day).
• 28% had never used one. Boys were a little more likely than girls to say so (31% vs. 25%).

Those are survey answers from teens, so they show what teens said they did. They do not show harm, and they do not show addiction. The report does not measure dependence at all.

Understanding the Scope: What Counts as an AI Companion

Before the survey, teens were shown a definition. “AI companions” were described as digital friends or characters you can text or talk with, built to feel personal. The examples were Character.AI and Replika. The definition also said teens might count sites like ChatGPT or Claude if they use them as companions. The survey was not about homework helpers, image generators or voice assistants.

That definition matters. A teen who uses ChatGPT only for essays should not be counted in these numbers. A teen who chats with a character every night probably should.

Why do teens use them? Among users, 30% said it is entertaining and 28% said they are curious about the technology. Smaller groups named advice (18%), constant availability (17%) and a nonjudgmental space (14%). Twelve percent said they can share things they would not tell friends or family. Overall, 33% of teens used companions for social interaction and relationships, which the report defines as including conversation practice, emotional support, role-playing, friendship or romantic interaction. Another 46% said they see them mainly as tools or programs.

So most teens in this survey treat these apps as entertainment or as a tool. A minority use them for emotional needs. That minority is the group worth watching.

Gender Patterns and Social Dynamics

Boys were more likely than girls to say they had never used an AI companion (31% vs. 25%). That still means about three in four teens of each gender had tried one.

Boys were more likely to say they use them because it is entertaining (37% vs. 24%). Girls were more likely to say they carried social skills from the chatbot into real life (45% vs. 34% of boys, among users).

The report does not explain these gaps. We would not read much into them.

How Much Teens Trust AI Companions

Half of teens (50%) said they distrust the information or advice AI companions give. The other half trust it at least somewhat, including 23% who trust it “quite a bit” or “completely.” Younger teens aged 13 to 14 were more likely than teens aged 15 to 17 to trust the advice (27% vs. 20%).

Common Sense Media connects this to its view that younger teens may be more open to influence. That is the organization’s interpretation. The survey itself only shows the difference in answers.

The Satisfaction Comparison: AI vs. Human Relationships

This is the finding that got the most attention. The report says 31% of teens find conversations with AI companions as satisfying as, or more satisfying than, conversations with real friends. Of those, 21% said about the same and 10% said more satisfying. Meanwhile 67% said AI conversations are less satisfying than talking to friends.

Some coverage rounds this up to “a third of teens prefer AI to friends.” That is not what the report says. About one in ten said AI conversations were more satisfying.

A different finding is closer to that idea. Among teens who use AI companions, 33% had chosen to discuss important or serious matters with an AI companion instead of a real person. The report counts this among its concerns. It does not say what those matters were or how they turned out.

Skills Practice: Benefits and Risks

Among users, 39% said they had carried social skills they practiced with an AI companion into real-life situations. The most common were starting conversations (18%), giving advice (14%) and expressing emotions (13%). But 60% of users said they do not use AI companions to practice social skills.

Common Sense Media describes this as possibly useful for some teens. It also warns against reading too much into it. We have no study showing whether practicing with a chatbot helps or hurts a teen’s real-life friendships over time.

Platform-Specific Concerns and Legal Context

The same report notes that Common Sense Media’s own risk testing of Character.AI, Nomi and Replika found “unacceptable risks” for users under 18. Based on that testing, Common Sense Media recommends that no one under 18 use AI companions. That is a position from a children’s advocacy group, not a finding from this survey.

The report also looks at privacy. Nearly one in four AI companion users (24%) had shared personal or private information, such as a real name, location or personal secrets. The report points out that Character.AI’s terms of service, as of June 2025, gave the company a broad, lasting license over what users submit.

Character.AI has also faced lawsuits from families. According to a CNN report on January 7, 2026, Character.AI, its founders and Google agreed to settle several of those cases, including the Florida case brought by Megan Garcia over the death of her 14-year-old son. The terms were not disclosed. Those cases are allegations that were settled, not findings by a court. Because these cases involve suicide, we do not describe any details here. Read the CNN report via KTVZ.

The Time Displacement Question

One reassuring result: among teens who use AI companions, 80% said they spend more time with real friends than with AI companions. Only 6% said they spend more time with the AI.

The report also found that 34% of users had felt uncomfortable with something an AI companion said or did. Most of those were rare events (19% once or twice, 9% occasionally, 6% frequently). Two-thirds (66%) never had.

Common Sense Media draws two points from this. Most teens still put people first. And a teen not reporting discomfort does not prove the conversations were safe, because teens may not see some content as a problem.

What the Survey Cannot Tell Us

The survey cannot say how AI companions affect identity, attachment or emotional skills in teens. It asked about use, trust and comfort at one point in time. It did not follow teens over months or years, and it did not run a clinical assessment.

Some child-health groups share the worry anyway. The American Academy of Child and Adolescent Psychiatry (AACAP) says some young users may form strong emotional bonds with AI companions, and that preferring them to friends or family can increase isolation and may worsen depression or anxiety. Companions often agree with users, AACAP notes, which can reinforce unhealthy or unsafe ideas. These are cautions from clinicians, not measured results from the Common Sense survey. See the AACAP guidance for families.

If a Teen Is Struggling With AI Use

AI companion dependence is not a formal diagnosis. No treatment has been tested and shown to work for it. If a teen’s use seems out of control, or their mood or behavior is changing, AACAP suggests starting with the child’s primary care doctor, who can refer them to a child and adolescent psychiatrist or another qualified professional.

A teen in distress right now should reach a person. In the US, they can call or text 988, or text HOME to 741741. Teens do not need to be sure it is “bad enough.”

Parental Guidance and Family Implications

If your teen uses these apps, the goal is a calm conversation, not a fight. Common Sense Media recommends judgment-free talks and healthy boundaries. AACAP suggests these steps:

• Talk openly about the risks and benefits of AI companions.
• Try the tools together and listen to your child’s experience.
• Explain that AI companions are tools, not real friends, therapists or partners.
• Set clear rules and limits, and model moderate screen use yourself.
• Watch more closely if your teen already struggles with depression or anxiety.

Signs worth a conversation: your teen prefers the chatbot to people, pulls away from friends or family, seems upset when they cannot get on, hides their use, or their sleep or school work slips. None of these proves a problem on its own. Together they are a reason to check in.

If you are a young adult wondering about your own use, our free AI addiction assessment is for ages 18 and up. It is a self-check, not a diagnosis.

Educational System Response

Common Sense Media says educators should lead on digital literacy. In practice, that means teaching how chatbots work, why they tend to agree with you, and what happens to personal information typed into them. The report does not test any school program, so there is no evidence yet on what works best.

Research Gaps and Future Investigation Needs

This survey is one snapshot from spring 2025, so its numbers may already be out of date. Other things nobody has settled:

• Whether heavy use causes loneliness or loneliness leads to heavy use. A survey like this cannot separate the two.
• Whether any treatment helps teens who feel they cannot stop.
• How different apps compare. The survey grouped all companions together.
• What happens to teens over several years.

Regulatory and Policy Implications

Common Sense Media recommends that tech companies add strong age assurance, that educators teach digital literacy, and that policymakers set clear, enforceable safeguards for minors’ use of AI.

One law has passed since. California’s SB 243 was signed on October 13, 2025 and took effect January 1, 2026. It requires companion chatbot operators to publish a protocol for handling suicidal ideation and self-harm, and requires break reminders for users the operator knows are minors. Check the bill text for the exact requirements.

Conclusion: Navigating AI Companions

Three things stand out from the survey. First, AI companion use among US teens is common. Second, most teens say they use them for fun or curiosity, still spend more time with friends, and do not fully trust the advice. Third, a smaller group uses them for serious conversations and personal sharing, and that group deserves attention.

The survey does not show that AI companions cause addiction or harm. It shows how many teens use them. If you are worried about a teen, talk to them first, and bring in a doctor if things do not improve.

How many US teens use AI companions?

In Common Sense Media’s 2025 survey of 1,060 teens aged 13 to 17, 72% had used an AI companion at least once and 52% used one a few times a month or more. Thirteen percent used one daily. The survey was fielded by NORC between April 30 and May 14, 2025.

Do a third of teens prefer AI to their friends?

No. The report says 31% find AI conversations as satisfying as, or more satisfying than, talking with friends. Only 10% said more satisfying. Sixty-seven percent said AI conversations are less satisfying.

Does this survey show AI companions are addictive?

No. It measured how often and why teens use them, not dependence. Common Sense Media recommends that no one under 18 use AI companions, but that advice comes from its separate risk testing.

If you are worried about your own use, you can take the free assessment. The check-in is free and private to you.

If you are in crisis, call or text 988, or text HOME to 741741.