DIRTY STATIC / DISPATCH Culture18 NOV 2026 • 10 MIN READ
A young person sitting alone on a concrete bench in a dark concrete hall, facing three empty concrete chairs, a phone beside them glowing caution yellow with a chat bubble, hazard stripes on the floor.
CultureINTELLIGENCE FOR THE MACHINE AGE

One Third Of Youth Use AI Companions

A Gallup and Boston Children's Digital Wellness Lab study of 15,925 young people aged 15 to 22 in 16 countries found a median of 33 percent used AI as a companion or friend. This feature follows where that use sits beside in person friendship, why two other global studies land on almost the same share, what the loneliness link does and does not show, and why young people still say they want a face in front of them.

Start reading ↓

Friendship used to need a second person in the room. A new global study suggests that for a large share of young people, it now sometimes needs only a phone. Gallup and the Digital Wellness Lab at Boston Children’s Hospital surveyed 15,925 people aged 15 to 22 across 16 countries between May 28 and August 21, 2026. A median of 33 percent said they had used AI as a social companion or friend.

That number alone could mislead in either direction. It does not mean a third of youth have replaced people with software. It also does not mean the use is trivial. The same study found that 93 percent still report a close in person friend, and that those friendships rate highest for strength and support. The companion sits beside the friendship in most lives. The open question is what it does once it is there.

The Bigger Story

Companionship was not the main reason young people opened AI at all. In the Gallup study, 82 percent had used AI to look up information and 68 percent had asked it for advice. Companionship came third at 33 percent. It grew with habit. Among people who used AI every day, the share who used it for company rose to 50 percent. Among people who used it less than once a week, the share was 24 percent.

The pattern also tracked the shape of a social life. Among young people with a close online only friend, 42 percent used AI as a companion. Among those without one, the share was 26 percent. In the United States the overall companion share was also 26 percent, lower than the global median. Young people in the United States, Canada, and the United Kingdom reported the least daily AI use of the countries studied, with only 20 percent using it once a day or more.

How young people said they used AI

Looked up information82%
Asked for advice68%
Used AI as a companion33%

Share of respondents aged 15 to 22 in the Gallup and Boston Children's Digital Wellness Lab study, median across 16 countries, surveyed May to August 2026.

Quality ratings complicate the picture further. Most young people who rated an AI companion gave it a medium rating, neither high nor low. The spread by country was wide. In Australia, about 45 percent of companion users rated the relationship as high quality. In Poland, only 12 percent did. Young people may not expect the same thing from a chatbot that they expect from a person, and the researchers warn that the ratings should be compared with care for exactly that reason.

A young person sitting on a bed at night in a concrete room, looking at a phone glowing caution yellow, an empty desk chair nearby

Context

Gallup is not the first study to land on one third. It is the third in a row, from a different team, in a different sample, using a different question.

Young Life, a Christian youth ministry, released its RELATE Project 2026 in September. It surveyed 7,105 people aged 13 to 24 across 11 countries between November 2025 and January 2026, and followed the survey with focus groups. There, 35 percent said they have an AI companion. Among the youngest group, aged 13 to 15, the share jumped to 45 percent. Among those aged 22 to 24 it was 29 percent. Young men reported a companion a little more often than young women, at 38 percent to 32 percent.

Common Sense Media found the same share a year earlier in the United States alone. Its 2025 report, Talk, Trust, and Trade Offs, drew on a survey of 1,060 teens aged 13 to 17 run by NORC at the University of Chicago. There, 72 percent had used an AI companion at least once and 52 percent used one a few times a month or more. And 33 percent used companions for social interaction and relationships, a bucket that covered conversation practice, emotional support, role play, friendship, and romance. Among teen companion users, 33 percent had also chosen the AI instead of a real person for an important or serious talk at least once.

Three studies do not prove one simple story. Their definitions differ, their age bands differ, and none of them watched behaviour directly. What they do together is harder to wave away. Across countries, ages, and question wording, roughly one in three young people now describe some part of their social life as happening with AI. We saw an early version of this shift in Students Confide In AI Before People, where the first telling of a feeling went to a chatbot because it felt less risky than a person.

Different Perspectives

The worried reading starts with loneliness, and the Gallup numbers give it real weight. Young people who used AI as a companion had 70 percent higher adjusted odds of being in the higher loneliness group. That is a statement about odds in a statistical model, not proof that a companion makes a person 70 percent more lonely. Gallup states the limit plainly in its own report. The data cannot show whether lonely young people turn to AI, whether AI companionship shapes their social experience, or whether other circumstances explain both.

whether young people are turning to AI because they feel lonely, or whether AI companionship might reinforce those feelings, remains an important question for future research
Andrew Dugan, Gallup, speaking to Wired

David Bickham, research director at the Digital Wellness Lab, made a related point in the same reporting. Feelings of loneliness, he said, may not be fully compensated for by online relationships. The wider Gallup findings support that caution without turning it into a verdict. Having a close in person friend was linked to higher life evaluation and lower loneliness. Having an online only friend was not linked to higher life evaluation, and people with an online only friend reported more loneliness regardless of how good they said that friendship was.

Young people themselves give a more mixed account, and it deserves to be heard without editing it into panic. In the Young Life study, six in ten companion users said they feel more comfortable sharing problems with AI than with some friends or family. One young woman in a focus group described what AI offers as conversations based on empathy, free of judgment. That appeal is easy to understand. A companion does not gossip, does not get distracted, and does not repeat what it was told at school the next day.

Yet the same Young Life sample said, clearly, that AI is not what it wants most. Sixty four percent said their preferred way to connect with close friends is in person, up from 54 percent in 2023. When a hard decision arrived, 70 percent said they would rather talk to a trusted person than ask AI or search online. Twelve percent said they cope with stress by talking to an AI companion, and nearly one in ten would turn to AI first for advice on something difficult. Those are real minorities. They are not the majority story.

The stricter critics start from age rather than from averages. Common Sense Media, in a separate risk assessment, rated social AI companions as an unacceptable risk for people under 18 and recommended that no one under 18 use them. That position sits in tension with the Gallup picture, where most young users also keep strong in person ties and rate the companion itself as middling. Both can be partly true at once. A product can be ordinary for most users and still be a poor place for a younger or more vulnerable user to take a serious problem.

Why It Matters

Start with what did not change. In Gallup, the factors tied most strongly to how young people rate their lives were all offline. Feeling that a household is getting by, feeling safe at home, and having a good relationship with a primary caregiver mattered more than any online habit measured. A young person with all three advantages rated life 3.4 points higher on a 0 to 10 scale than a young person with none of them. No app feature in the study came close to that gap.

That context matters because it keeps the companion in proportion. AI companionship is now common enough to count as a social fact, not a fad among a few heavy users. It is also, for most of the young people surveyed, one thread in a social life that still runs mainly through people they can sit with. The risk is not that friendship has ended. The risk is quieter. When company is always available, always patient, and never in need of anything back, the harder work of ordinary friendship can start to feel like a flaw in people rather than a part of it. We traced a version of that bargain in Your AI Girlfriend Has a Changelog, where the partner users loved could change overnight because the partner was a product.

A group of teenagers sitting on concrete steps talking to each other, their phones face down beside them with a faint caution yellow glow

For parents, schools, and youth workers, the useful signal is not the headline share by itself. It is the pattern underneath it. Daily use, online only friendship, and a lack of a trusted adult outside the family all clustered with heavier companion use and with more loneliness in these studies. Young Life found that four in ten young people are not fully confident they have an adult who is genuinely interested in listening to them, and that the average number of close relationships fell from 5.3 in 2023 to 4.4 in its latest study. A companion that fills a gap is telling you where the gap is.

Closing Reflection

There is a temptation to read one third as a verdict on a generation. The studies resist that reading if you stay with them for more than a headline. Most young people still have close in person friends. Most still prefer a face when it counts. Many rate the AI in their pocket as merely medium, a thing that is there, useful at odd hours, and not confused with the real thing.

What has changed is the floor. Company used to require another person to be awake, willing, and free. Now a passable version of being heard is available at any hour, to anyone with a phone, including a 14 year old who does not yet have an adult they trust. That will not replace friendship for most people. It may still reshape what the loneliest hours look like, and who, or what, gets trusted inside them. The question these three studies leave open is not whether young people can tell AI from a friend. Most of them can. It is whether the adults and institutions around them will close the gap the companion is currently filling, before filling it becomes the habit that stays.

Reporting draws on the Gallup and Boston Children’s Digital Wellness Lab report Friendship in an Online Offline World, surveyed May 28 to August 21, 2026 and released November 2026, Young Life’s RELATE Project 2026, and Common Sense Media’s Talk, Trust, and Trade Offs from 2025, checked November 2026.

Sources

■

END OF TRANSMISSION

Stay curious. Question everything.

CONTINUE THE DESCENT

MORE FROM DIRTY STATIC ↘