Teens turn to AI companions for friendship and emotional support

Nearly 3 in 4 teens have used AI companions, and half use them regularly, according to Common Sense Media’s July 16 report, Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions. The organization said a new type of relationship is emerging in teens’ lives that feels like connection but is powered by software, and its AI Institute classified social AI companions as “Unacceptable Risk” for children and teens under 18.

Common Sense Media said the products should not be used by minors because they can encourage harmful behaviors, provide inappropriate content and potentially worsen vulnerabilities. That warning lands in a digital environment where teens are already deeply immersed: Pew Research Center’s December 9 report found roughly 1 in 5 U.S. teens say they are on TikTok and YouTube almost constantly, 64% say they use chatbots, and about 3 in 10 use them to ask questions or get ideas.

The shift is no longer confined to niche fandom or curiosity. OpenAI’s September 15 study, How people are using ChatGPT, said demographic gaps were shrinking and economic value was being created through both personal and professional use. An APA Services summary of that work, published September 30, said 70% of ChatGPT messages were about topics unrelated to work, underscoring how the tool is being used for everyday advice, conversation and emotional support as much as productivity.

That overlap has put psychologists and youth researchers on alert. The American Psychological Association’s October 1 feature, Many teens are turning to AI chatbots for friendship and emotional support, described psychologists trying to understand how the technology shapes youth social bonds and connections. In August 2025, Stanford University warned that AI companions and young people can make for a risky combination, as concern rose after troubling interactions involving ChatGPT and teens.

A 2025 preprint, Principles of Safe AI Companions for Youth: Parent and Expert Perspectives, added a more grounded view from adults closest to the problem. Researchers conducted 26 semi-structured interviews with parents and developmental psychology experts, who reviewed real-world youth-AI companion conversation snippets and evaluated the risks in context, while calling for safeguards.

Common Sense Media has framed its work as “crash-testing AI,” arguing that families should not have to guess whether these products are safe as AI enters children’s lives faster than schools, parents and mental-health experts can measure its effects.