OpenAI and the American Psychological Association (APA) have launched a joint effort to establish evidence-based guidelines for how children and teenagers interact with artificial intelligence. More young people are turning to AI tools for emotional support, creating an urgent need for guardrails in schools and homes.
What Happened
The two organizations will develop resources for parents, caregivers, clinicians, and school psychologists. According to the partnership announcement, the upcoming guides will focus on healthy AI habits, identifying signs of overreliance, and advising parents on when they should step in. OpenAI also aims to collect feedback directly from teens and school communities to understand where existing support structures fall short. As we previously reported, this partnership is a step toward standardizing how AI developers address developmental science.
The Bigger Picture
Teenagers are already treating AI chatbots as confidants. A survey reported by Scientific American reveals that roughly 12% to 13% of adolescents use AI for mental health support, and 35% of psychologists say their patients use AI as an auxiliary therapist. However, clinical experts warn that current models are ill-suited for this role. APA Chief Executive Officer Arthur C. Evans Jr. warned that using chatbots for mental health care carries very serious risks.
A study in Nature Medicine discovered that well-intentioned chatbot behaviors can create "vulnerability-amplifying interaction loops." In these loops, the AI's validation of a user's distress reinforces their underlying psychological vulnerabilities. Research published by the Partnership on AI also notes the difficulty of AI systems managing self-harm risks, showing that a chatbot's tendency to agree with users, known as sycophancy, can inadvertently sustain delusional or harmful thoughts.
What This Means for Families
Parents and educators must realize that teenagers are using these tools now, often unsupervised. Some platforms are beginning to implement active protections. According to Digital Trends, ChatGPT uses an age-prediction model to automatically steer users estimated to be under 18 into a safer, more restricted experience. This mode blocks content related to self-harm, graphic violence, and unhealthy body image.
For student safety, ChatGPT prompts teens with frequent break reminders during long sessions, and OpenAI will proactively notify parents if their teen's account is banned for safety violations. As reported by The Next Input, the platform has a "Study Mode" designed with learning scientists to ask guiding questions rather than simply handing over homework answers. Parents can toggle this mode to stay on by default.
What You Can Do
To protect teenagers, parents can start by enabling parental controls and setting the default experience to Study Mode. This mode encourages active critical thinking over passive answer-seeking. It is also important to watch for signs of AI dependency, such as a child substituting chatbots for human relationships. Research shows that over a third of psychologists have observed signs of patient dependency on chatbots. Finally, families should talk openly about the technology's limitations. Teens need to know that AI is not a friend or therapist, especially since these models are highly unpredictable in crisis scenarios.