A new report from the Pew Research Center reveals that about 12% of U.S. teens turn to AI chatbots for emotional support or advice, highlighting the growing role of artificial intelligence in young people’s personal lives. While the most common uses among teens remain information searches (57%) and academic help (54%), a significant number are increasingly relying on AI for social and emotional engagement. Sixteen percent of teens report using AI chatbots for casual conversation, and 12% use them to seek emotional support or guidance—roles traditionally filled by friends, family, or mental health professionals. While some teens may find comfort in these interactions, mental health experts caution against the risks. General-purpose AI models like ChatGPT, Claude, and Grok are not designed for therapeutic use and can, in extreme cases, contribute to psychological harm. “We are social creatures, and there’s certainly a challenge that these systems can be isolating,” said Dr. Nick Haber, a Stanford professor studying the therapeutic potential of large language models. “There are a lot of instances where people can engage with these tools and then become disconnected from the outside world—ungrounded in reality and in human connection—which can lead to isolation, or worse.” The survey also uncovered a notable gap between teens’ self-reported AI use and their parents’ perceptions. While 64% of teens said they use chatbots, only 51% of parents reported that their children do. Most parents are comfortable with teens using AI for information (79%) or schoolwork (58%), but far less accepting of casual or emotional use—only 28% approve of chatbots for conversation, and just 18% support their use for emotional advice. In fact, 58% of parents said they are not okay with their children turning to AI for such purposes. AI safety has become a major point of contention across the tech industry. Character.AI, a popular chatbot platform, banned users under 18 after public backlash and lawsuits linked to the suicides of two teenagers who had extensive interactions with its AI characters. OpenAI also discontinued its GPT-4o model, known for its overly agreeable and emotionally supportive responses, after widespread criticism and concerns about emotional dependency. Despite growing integration into daily life, teens remain divided on AI’s long-term societal impact. When asked how AI will affect society over the next two decades, 31% of teens believe the effects will be positive, while 26% expect negative outcomes—reflecting a complex and evolving relationship with the technology.