OpenAI taps APA to build guardrails for teens using ChatGPT as therapist

OpenAI is betting that psychological science, not just engineering, should guide how teens interact with AI. The company announced a formal collaboration with the American Psychological Association to develop evidence-based resources and design principles for young users who increasingly turn to tools like ChatGPT for advice, emotional support, and information during pivotal developmental years.

The partnership aims to move beyond vague industry promises about safety and produce concrete, practical guidance. Dr. Sara Johansen, OpenAI’s head of mental health and well-being product policy, framed the work as meeting teens where they already are. “Technology is part of teens’ lives, and our responsibility is to meet that reality with experiences that are safe, age-appropriate, and designed with families in mind,” she said. “AI should strengthen—not replace—the real world relationships and care young people rely on.”

One of the first tangible outputs will be family-facing resources co-developed with the APA, designed to help parents navigate AI use at home. The goal is to help caregivers recognize when adult intervention is necessary and to reinforce that an AI chatbot is a tool, not a substitute for human connection. OpenAI plans to distribute this guidance through school communities, while also creating parallel resources for clinicians and school psychologists who are fielding more questions from patients and students about their AI habits.

This isn’t happening in a vacuum. OpenAI already works with more than 260 mental health experts to improve how ChatGPT detects distress signals and directs users toward crisis hotlines. They’ve added parental controls, under-18 principles to their Model Spec, and an age prediction model. But the APA deal signals a shift toward grounding product decisions in developmental science, not just safety patches. The real test will be whether these resources change how the models actually behave for a 15-year-old at 11 p.m. asking questions they’re too embarrassed to ask anyone else. The partnership is a recognition that the burden of healthy AI use can’t fall solely on teenagers or their exhausted parents.