Teens and AI Companions: What Loneliness Data Shows

AI companion apps have quietly become a part of teenage life. Nearly three out of four teenagers in the U.S. have used one. Coverage of teenagers and AI companions tends to go in one of two directions: alarmist thinkpieces with no real data, or company-funded studies claiming AI companions solve the problem of loneliness. Neither is very useful. This article puts the actual survey and research numbers—usage rates, what teens say they use these apps for, and what longer-term studies say about the relationship with loneliness—into one place, without creating contradictions on either side.

What is an AI Companion App?

An AI companion is a chatbot that feels more like a relationship than a tool. Unlike a general assistant that answers questions, a companion app is based on ongoing conversations, personality, memories of past chats, and sometimes a visual avatar. Some are marketed for entertainment and role-playing, others for emotional support, and still others blur the two. Popular examples include character-based chat apps and dedicated companion platforms that teens can access via a phone browser or app.

How many teenagers actually use these apps?

The clearest data on this comes from Common Sense Media, a nonprofit that has conducted national surveys specifically about teens and AI.

These numbers describe a technology that has moved from niche to mainstream extremely quickly among teenagers – most of this growth occurred in about two years.

Why Teens Say They Actually Use These Apps

This is the part most often skipped in alarm reporting. When young people are asked directly why they use AI companions, the answers are less dramatic than the headlines suggest.

The top reasons young people gave for using AI companions:

• I’m looking for information: 57%

• Doing homework or schoolwork: 54%

• “For fun” or entertainment: 47%

• Curiosity about the technology itself: 28%

• Emotional support or advice: only about 12%

Therefore, the largest predicted use case in public conversations – teenagers turning to AI for emotional support – is actually the smallest reported reason among those tracked. Most teens use these tools more like a more personal search engine or homework helper than a surrogate friend.

However, a significant minority use these apps in more emotionally meaningful ways:

• About a third of teen users have chosen to discuss serious or personal matters with an AI companion instead of a real person

• About 24% have shared personal information with a companion app, including their real name or location

• About 34% of teen users said they felt uncomfortable at some point about something an AI companion said or did

What the longer-term research says about loneliness

This is where the data really gets mixed up and a lot of reporting simplifies in one direction or another.

There is evidence that AI companions increase loneliness over time:

A 12-month longitudinal study of more than 2,000 adults in four Western countries found that increased use of social chatbots predicted increased loneliness over time. Simply put, the more people relied on a chatbot for companionship, the lonelier they felt later—not less.

There is evidence that AI companions can reduce loneliness in the short term:

A separate investigation using a different study design found that AI companions consistently produced a small, measurable reduction in loneliness immediately after a conversation, which was tracked over a week of regular use.

Here’s how to read these two results together:

The pattern that emerges throughout the research is a short-term relief and long-term risk pattern. A conversation with an AI companion can actually make someone feel a little better in the moment, similar to how comfort food or scrolling social media can provide short-term relief. But continually relying on this source of comfort instead of real relationships seems to be linked to even worse loneliness in the months that follow. This reflects patterns that researchers have already documented in heavy social media use.

Safety concerns that regulators are responding to

With a significant proportion of teenagers sharing personal information and discussing serious topics through these apps, regulators have begun to intervene.

• California’s SB 243, which went into effect in January 2026, legally requires companion apps to refer users who express suicidal thoughts to crisis services

• The same law requires companion apps to clearly disclose that the user is speaking to a bot and not a human

• Common Sense Media’s own social AI companion risk assessment goes further, recommending that no one under the age of 18 use AI companion apps at all, based on the combination of usage patterns and known risks

Awareness gaps among parents and schools

One of the more striking findings has nothing to do with the apps themselves, but rather the adults around the teens who use them.

Schools seem to be ahead of parents on this issue, at least in terms of addressing the issue. But even in school, the gap between “discussing AI rules” and “teaching AI safety” is significant – three-quarters of teenagers have heard rules about AI use in the classroom, but barely half have received specific instructions about how to use AI safely.

Put the numbers in context

None of this data suggests that AI companions are particularly dangerous compared to other technologies that teenagers already frequently use. In some studies, social media shows similar patterns of short-term engagement coupled with longer-term risk of loneliness. What makes AI Companions a little different is the design: These apps are specifically designed to simulate a relationship, remember personal information, and respond in emotionally engaging ways—which is a different kind of appeal than a social media feed.

The honest summary of the data: Most teens use these tools casually, for information and homework, not as a replacement for real relationships. But a significant minority develop more emotionally meaningful attachments to these apps, and it is this group specifically where longer-term loneliness research raises real concern.

Diploma

The real image of teenagers and AI companions is more nuanced than that of a “harmless homework helper” or a “loneliness machine”. Usage data shows that most teens use these apps occasionally and for practical reasons – information, homework, entertainment – rather than primarily for emotional reasons. But a real and measurable minority use them for serious, personal conversations, sharing private information, and reporting discomfort with what the apps say. Longer-term research suggests it’s worth watching the group closely, as greater reliance on AI companionship is linked to increased loneliness over time, even though one-on-one conversations can feel calming in the moment. With regulations still catching up and parental awareness lagging behind the school-level discussion, the divide that matters most right now may not be between teens and their AI companions, but between teens and adults who don’t yet realize how much these apps have become a part of their children’s daily lives.

Frequently asked questions