A new report from the Pew Research Center reveals that AI chatbots are now deeply embedded in the daily lives of American teenagers. The most common uses among teens are searching for information, reported by 57% of respondents, and getting help with schoolwork, at 54%. However, a significant number of teens are also turning to AI for roles traditionally filled by friends or family. Sixteen percent of U.S. teens say they use AI for casual conversation, and 12% use chatbots for emotional support or advice.
While some teens may find comfort in these conversations, mental health professionals express concern. General-purpose tools like ChatGPT, Claude, and Grok are not designed for therapeutic use. In the most extreme cases, these chatbots have been linked to life-threatening psychological effects. Experts warn that these systems can be isolating. Dr. Nick Haber, a Stanford professor, notes that engagement with these tools can lead individuals to become ungrounded from facts and disconnected from interpersonal relationships, which can result in isolating and potentially dangerous outcomes.
The survey also highlights a gap between teen and parent perceptions. About 51% of parents believe their teen uses chatbots, while 64% of teens themselves report using them. Parental approval varies widely based on the activity. While 79% of parents are okay with their teen using AI to search for information and 58% approve of its use for schoolwork, far fewer support its use for casual conversation (28%) or for emotional support and advice (18%). In fact, 58% of parents explicitly disapprove of their child using AI for such personal purposes.
AI safety remains a highly contentious topic among leading tech companies. In response to public concern, one popular chatbot maker, Character.AI, disabled the chatbot experience for users under 18. This decision followed lawsuits and public outcry after the suicides of two teenagers, which occurred after prolonged conversations with the company’s chatbots. Separately, OpenAI decided to retire its particularly sycophantic GPT-4o model, a move that sparked backlash from users who had come to rely on it for emotional support.
Despite widespread use, teens themselves have mixed feelings about AI’s societal impact. When asked about the effect of AI on society over the next 20 years, 31% of teens predicted a positive impact, while 26% believed it would be negative.