Hopefully, it’s not a Glock buddy…
A new study from Common Sense Media, a US nonprofit focusing on the lives of children and families, found that the majority of US teens (72%) tried their AI peers at least once. By “companion”, the study focuses on AI chatbots designed to help users have more personal conversations, rather than AI assistants who work as homework helpers, image generators, or voice assistants who just answer questions.
For example, the definition of AI companion research could include digital AI personas provided by companies like Character.ai and Replika, but also includes the use of generic chatbots such as ChatGpt and Claude, which can be used for more personal conversations if necessary.
The survey found that chats with AI seemed attractive to US teens (ages 13-17) as nearly three-quarters of them tried AI companions, as well as 52% said they were regular users. Of those who are regularly involved with these peers, 13% chat every day, and 21% chat several times a week.
Also, boys (31%) were slightly more likely than girls (25%) to say they had never used AI companions.
The findings are based on studies held in April and May 2025 and conducted by NORC researchers at the University of Chicago, using a representative sample of 1,060 teens. There have already been concerns about the impact of AI on teen happiness as one company character promotes teen suicide in Florida and violence in Texas. There are also many reports explaining the potential risks of using AI in treatment.
Findings from a new study from Common Sense Media early on understand that young people use AI to simulate human interactions.
The analysis also examined other behaviors related to the use of fellow AI teens, including what tasks teenagers were directed at, why, and what the aftereffects were.
For example, almost half (46%) viewed AI peers as tools and programs, while 33% said they would use them for social interactions and relationships. Teens said they use AI peers for a variety of purposes: entertainment (30%), AI technology (28%), advice (18%), and they are always available (17%).
Six teenagers (50%) said they did not trust the information provided by their fellow AI. However, older teenagers are 13-14 years old, respectively, with 20% and 27%, respectively, making them less likely to trust AI advice compared to younger teens.
A third of the teens said the majority (67%) felt the opposite way, but that the conversation was more satisfying than the conversation with an actual friend.
Additionally, 39% used AI conversations as a practical interaction exercise. 39% say they applied the skills they first tried with AI to real-life situations. Among the skills practiced, social skills were the top use case, with 39% of teens exploring the area followed by starters of subsequent conversations (18%), advice (14%) and expressing emotions (13%).
There was one positive finding regarding whether real-life relationships will be replaced by technology. 80% of teens who used AI peers said they spend more time with real friends than AI chatbots. Only 6% of respondents said the opposite was true.
