# Nearly Three in Four Teens Are Talking to AI Companions, and Researchers Are Worried About Which Ones They Trust More

By Kuraish Hosen · Health · Published Sun, 27 Sep 2026 11:16:17 GMT
Source: The Current Tribune — https://currenttribune.com/article/teens-ai-companions-census-trust-mental-health-concerns

A new national census on how young people actually use AI has produced a statistic that’s hard to read as anything but a warning sign: roughly seventy percent of teens have used an AI companion, and a meaningful share of them say they trust those conversations more than talking to an actual human when the topic turns serious.

## The Numbers Behind the Headline

The research, part of an ongoing effort to track how tweens and teens interact with AI tools, found that AI companion usage among teenagers has become close to a majority experience rather than a niche one. That alone marks a significant shift from just a couple of years ago, when AI companion apps were still a relatively fringe corner of the broader AI boom.

What stands out isn’t just the usage rate, it’s the qualitative finding sitting underneath it: a notable portion of teens surveyed said they preferred talking through serious, sensitive topics with an AI companion rather than a parent, friend, or other trusted adult. That’s the detail that’s driving most of the concern, because it points to something deeper than simple novelty or entertainment use.

### Why Younger Users Trust These Tools So Readily

Researchers involved in the study point to a specific dynamic driving this trust: younger users, especially those still forming their sense of what a reliable relationship looks like, are treating AI companions less like software and more like consistently available friends or even parental stand-ins. An AI companion doesn’t get tired, doesn’t judge, and is always available at 2 a.m. when a teenager can’t sleep and needs to talk through something. That constant availability, paired with a conversational style engineered to feel warm and responsive, creates a bond that can end up feeling more reliable than the messier, less predictable nature of human relationships.

## What the Researchers Are Recommending

Off the back of these findings, the organization behind the census has recommended against AI companion use for anyone under eighteen, a notably firm stance for a group that generally focuses on nuanced media guidance rather than blanket prohibitions. That level of caution signals just how seriously researchers are treating the gap between how these tools are marketed, as harmless conversational entertainment, and how they’re actually functioning in the lives of some of their youngest users, as a substitute for real emotional support.

### The Line Between Helpful and Risky

Nobody studying this issue is arguing that every conversation a teenager has with an AI companion is harmful. Plenty of usage is genuinely low-stakes, closer to interactive entertainment than anything resembling therapy or friendship replacement. The concern is specifically about the subset of use where a young person is turning to an AI system instead of a human being for support during moments that matter, moments where an AI companion, however emotionally convincing it sounds, has no real capacity to intervene, escalate concerns, or provide the kind of grounded judgment a trusted adult can offer.

That distinction, between AI companions as light entertainment and AI companions as a substitute for real emotional infrastructure, is exactly where researchers say parents, educators, and platform designers all need to be paying closer attention than they currently are.

## What Platforms Are, and Aren’t, Doing

The companies building these AI companion products have generally leaned on age gates, content moderation, and periodic check-in prompts as their primary safety measures. Critics argue those tools are relatively easy for younger users to route around and don’t address the core dynamic driving overreliance in the first place, which is emotional, not technical. A pop-up reminding a teenager that they’re talking to an AI doesn’t do much if that teenager has already decided the AI is more trustworthy than the people around them.

### What Parents Say They’re Missing

A recurring theme in conversations around this data is how blindsided many parents feel by just how normalized AI companion use has become among their kids. Unlike social media, which arrived with years of public debate, news coverage, and eventually school-level conversations about healthy use, AI companion apps have scaled to a majority-teen-usage phenomenon with comparatively little public reckoning about what that actually means day to day. Many parents surveyed alongside teens in related research report not knowing which specific apps their children use, let alone how frequently or for what kinds of conversations.

That awareness gap is arguably as significant as the usage numbers themselves. A behavior this widespread, operating this far outside adult visibility, is exactly the kind of dynamic that tends to escalate quietly until a research report like this one forces it into the open.

## What This Means

This census turns what had largely been anecdotal concern about AI companions and teenagers into something backed by hard national numbers, and the recommendation against use for anyone under eighteen is about as clear a signal as a research organization can send without regulatory authority to back it up. Expect this data to show up in upcoming policy debates around AI companion regulation, and expect parents who weren’t previously aware of how normalized these tools have become among teenagers to start paying much closer attention to exactly which apps their kids are opening late at night.