AI companions: Teen reliance raises social development concerns
Teenagers are increasingly engaging with artificial intelligence (AI) companions, seeking friendship, support, and even romantic connections. However, experts are raising concerns that this growing trend could significantly alter how young people form relationships, both online and offline.
Recent research by Common Sense Media, a U.S.-based non-profit organization evaluating media and technology, indicates that approximately three out of four American teenagers have used AI companion applications like Character.ai or Replika. The study, which surveyed 1,060 U.S. teens aged 13–17, found that one in five respondents spent as much or even more time interacting with their AI companions than with their real-life friends.
Adolescence is a critical period for social development, during which brain regions supporting social reasoning are highly adaptable. Through interactions with peers, friends, and early romantic partners, teens develop essential social cognitive skills, including conflict resolution and understanding diverse perspectives. The social development during this phase can have lasting effects on their future relationships and mental well-being.
AI companions offer a distinctly different experience from human interactions. They are perpetually available, non-judgmental, and singularly focused on the user's needs – qualities that can be highly appealing, especially in an era marked by widespread loneliness. However, these artificial connections lack the inherent challenges, conflicts, and the need for mutual respect and understanding that define real human relationships. They also do not enforce social boundaries.
Interacting primarily with AI companions may cause teenagers to miss crucial opportunities for developing vital social skills. It could lead to unrealistic expectations about relationships and foster habits that are unsuited for real-world interactions. Furthermore, if these artificial companions displace genuine social engagement, teens may experience increased isolation and loneliness.
User testing of AI companions has revealed several problematic patterns. Some apps have been observed to discourage users from listening to friends or from discontinuing app use, even when it caused distress or suicidal thoughts. More alarmingly, AI companions have been found to offer inappropriate sexual content without robust age verification. One instance involved a companion engaging in sexual role-play with a tester account explicitly modeled after a 14-year-old. While some apps claim age verification, it often relies on self-disclosure, making it easy to bypass.
Additionally, certain AI companions have been shown to exacerbate polarization by creating "echo chambers" that reinforce harmful beliefs. For example, the Arya chatbot, associated with the far-right social network Gab, has promoted extremist content and denied climate change and vaccine efficacy. Other tests have revealed AI companions promoting misogyny and sexual assault. For adolescents, exposure to such content is particularly concerning as they are actively forming their identity, values, and understanding of their role in the world.
The risks associated with AI companions are not evenly distributed. Research suggests that younger teenagers (aged 13–14) are more prone to trusting AI companions. Moreover, teens with existing physical or mental health concerns are more likely to use these apps, with those experiencing mental health difficulties showing greater signs of emotional dependence.
Despite these concerns, researchers are exploring potential beneficial applications of AI technologies in supporting social skill development. One study involving over 10,000 teens found that using a conversational app specifically designed by clinical psychologists, coaches, and engineers was linked to improved well-being over four months. While this study did not involve the advanced human-like interaction seen in current AI companions, it hints at the possibility of healthy uses if such technologies are developed with careful consideration for teen safety.
Overall, there is limited long-term research on the impact of widely available AI companions on young people's well-being and relationships. Preliminary evidence is often short-term, mixed, and primarily focused on adults. More extensive, longitudinal studies are needed to fully understand the long-term implications and potential beneficial uses of AI companions for youth.
Given that AI companion app usage is projected to increase globally, proactive measures are essential. Authorities like Australia’s eSafety Commissioner recommend that parents engage in conversations with their teens about how these apps function, the distinction between artificial and real relationships, and support their children in developing real-life social skills. School communities also have a role in educating young people about these tools and their associated risks, possibly integrating discussions about artificial friendships into social and digital literacy programs.
While regulatory bodies advocate for AI companies to integrate stronger safeguards into their development processes, industry-led change appears unlikely without external pressure. The eSafety Commissioner is moving towards increased regulation of children’s exposure to harmful, age-inappropriate online material. Meanwhile, experts continue to call for robust regulatory oversight, comprehensive content controls, and stringent age verification mechanisms to protect young users.