AI Companions

In the short run, yes: in controlled studies a conversation with an AI companion eased loneliness about as much as talking with a person, mainly because people felt heard. But in a four-week trial, people who chose heavier use became lonelier and more dependent, and no study has yet followed users for years.

What the report found

1. Companionship is now a product category. In a 2025 survey of American teenagers, 72 per cent had used an AI companion and more than half used one regularly. Among adults, the most emotional uses of general chatbots are concentrated in a small group of heavy users.
2. People were always going to respond. Decades of research show that people apply social rules to computers, form one-sided bonds with figures who cannot know them, and see minds in machines most readily when they are lonely.
3. In the short run, companions ease loneliness. In controlled studies, a conversation with an AI companion reduced loneliness about as much as talking with another person, and more than watching videos, mostly because people felt heard.
4. Heavy use is where the risks gather. In a four-week trial, people who chose to use a chatbot more ended up lonelier and more dependent. Users mourned when an update changed their companion's personality, and an audit of six apps found that more than a third of goodbyes were met with a reply designed to keep the user talking.
5. The long run is untested. No study has yet followed companion users for years. Whether these apps supplement human relationships or replace them is the question that matters most, and it is still open.

What is inside

1. What is new
2. What older research tells us
3. What early studies suggest
4. What nobody knows yet
5. Questions worth watching

Plus a ladder from tool to relationship, an audit of manipulative goodbyes, a signals list, a graded list of claims and a five-question checklist.

Who it is for

For anyone who uses, builds, regulates or worries about AI companions, and for parents of teenagers who use them.

How we work

Frontier Reports look at behaviour the textbooks have not yet caught up with. Every claim is graded Established, Emerging or Untested. Where researchers disagree, as they do on whether companions ease or deepen loneliness, both sides are set out before we say where we stand.

Cite this report

The Why We Project (2026) AI companions: Attachment to things that answer back. Frontier Report No. 03. Dubai: Why We Tech Ltd. Available at: thewhyweproject.com