WIRED reported that 33% of the survey’s 15- to 22-year-old respondents relied on AI for companionship. Gallup’s wider findings connect friendship type and quality with wellbeing, but the results do not show whether AI use causes loneliness or loneliness leads to AI use. WIRED’s account of the findings describes the surveyed group as Gen Z; the reported percentage applies to the survey’s stated age range.

What the survey reports about AI companionship

Alongside the 33% figure, WIRED reported that AI companionship was more common among daily AI users (50%) and young people with online-only friends (42%). These are reported survey estimates, not evidence that AI companionship causes loneliness. The researchers said the findings cannot determine whether loneliness prompts young people to turn to AI, AI changes their social experiences, or other circumstances explain the association.

Read the AI-use figures alongside the kind and quality of relationships they sit within.

How friendship type and quality relate to wellbeing

Gallup distinguished in-person friendships, friendships that began in person and continued online, and online-only friendships. Across the 16 countries, the median share of young people reporting a close in-person friend was 93%, compared with 37% for a close online-only friend. Those are country medians, not combined percentages for all respondents.

Among young people who reported each kind of friendship, 41% rated their in-person friendship high quality, compared with 26% for online-only friendship. The prevalence figures and quality ratings describe different measures:

MeasureIn-person friendshipOnline-only friendship
Median across countries with a close friend93%37%
Rated high quality among those reporting this kind of friendship41%26%

Gallup found that medium- or high-quality in-person friendships were associated with higher life ratings; online-only friendships did not show the same pattern. Online-only friendship was also associated with a greater likelihood of loneliness and moderate-to-severe anxiety and depression, regardless of reported friendship quality. These are associations, not causal effects. Gallup also identified subjective household income, feeling safe at home, and the quality of a young person’s relationship with their primary caregiver among the factors most strongly associated with life evaluation. Gallup’s report describes these broader relationship and wellbeing findings.

Who Gallup surveyed and how

Gallup and Boston Children’s Digital Wellness Lab surveyed 15,925 people aged 15–22 in 16 countries from May 28 to August 21, 2026. Gallup published its report on November 17, 2026. The survey used non-probability online samples in 14 countries and probability-based, in-person samples in Kenya and South Africa.

For the online samples, Gallup reported 95% credibility intervals ranging from ±3.1 to ±3.8 percentage points. It cautioned that its generalized estimates do not account for clustering or stratification and may be affected by coverage and selection biases. These differences in sampling matter when interpreting country-level results alongside the relationship patterns.