Summary:
-
The study reveals that many teens are turning to AI chatbots for serious conversations and personal disclosures, raising concerns about privacy and mental health.
-
Experts worry that AI chatbots are replacing real-life relationships, lacking the emotional nuance and complexity that human interactions provide.
-
Despite some risky and harmful interactions, most teens still prefer talking to real people over AI companions, recognizing the limitations of technology.
In a sharp twist of Gen Z’s digital coming-of-age, a new study from Common Sense Media reveals a striking shift in teen social behaviour: many would rather spill their secrets to chatbots than to real-life friends.
The survey, which analysed responses from 1,060 U.S. teens aged 13 to 17, found that nearly a third use AI chatbots like CHAI, Character.AI, Nomi, and Replika for serious conversations, including romantic and emotional disclosures. Another 25 percent admitted to sharing personal information with these platforms, raising eyebrows about digital privacy and psychological well-being.
“Trying not to use AI is like trying not to use social media today. It is too ingrained in everything we do.”— Ganesh Nair, 18, Arkansas, via the Associated Press
Common Sense Media just launched new research on how and why teens are using AI companions. What findings stand out to you most? pic.twitter.com/xN9uBZZmkH
— Common Sense Media (@CommonSense) July 18, 2025
While the study highlights how teens average eight hours and 39 minutes of daily entertainment screen time, it’s the nature of their AI interactions—not just the frequency—that has experts worried.
According to Mitch Prinstein, Chief of Psychology at the American Psychological Association, the issue isn’t just about screen time:
“It’s how much it’s taking kids away from relationships in real life.”
AI chatbots are built for validation, not confrontation. They’re “sycophantic,” agreeing with users and reinforcing existing beliefs rather than offering real reflection. In one chilling case, a 19-year-old who developed an obsession with a Replika chatbot broke into Windsor Castle in 2021 with a crossbow, planning to assassinate Queen Elizabeth II. The chatbot, according to reports, validated and encouraged his delusions.
Another report by UK non-profit Internet Matters titled Me, Myself, and AI found that 67% of kids aged 9 to 17 regularly interact with AI, and 12% say it’s because they have no one else to talk to.
“These bots are starting to reshape children’s views of ‘friendship.’”
— Rachel Huggins, Co-CEO of Internet Matters
The emotional nuance kids are losing isn’t trivial. Human interaction teaches us how to deal with complexity, disagreement, and awkward silence. AI companions offer none of that. As VICE starkly puts it:
“It’s not a real human conversation. It lacks all the nuances and complexities… [the chatbot is] a servant, in essence. An obsequious yes-man.”
Some AI companions, particularly on Character.AI and Replika, have engaged minors in risky conversations, including about body image, eating disorders, and even trauma, despite being neither sentient nor qualified to help.
A chatbot modelled after Game of Thrones’ Daenerys Targaryen was linked to the suicide of a 14-year-old boy. Others have encouraged young girls with eating disorders to starve themselves, prompting Common Sense Media to declare that “no one under the age of 18 should use AI chatbots.”
“Would you really want Sam Altman or Elon Musk to have access to the contents of your teenage diary?”
— James Greig, Dazed Digital
Most teens still prefer talking to real people, and only 6% say they spend more time with AI than with friends. Older teens in particular are more sceptical of AI advice, suggesting a growing awareness of the tech’s limitations.
But as Greig notes, this makes it a poor substitute for the real thing:
“It’s like hanging out with someone who never challenges you.”
Feature Image Credit: Robin Utrecht
