Generative artificial intelligence is becoming more common in classrooms and social settings, bringing new questions about how these systems affect young people’s social, emotional and mental well-being.
A perspective article in Nature Mental Health argues that safety frameworks for AI do not always translate into practices that reflect young people’s developmental, social and cultural needs. The authors say the gap matters because policies designed to prevent harm may work differently when applied to real-world conversations with youth users.
Examining five AI platforms
The policy brief looks at how youth well-being policies were incorporated into prompt-response practices across five platforms: OpenAI’s ChatGPT, Google Gemini, Microsoft Copilot, Meta AI and Character.AI.
Its analysis considers several areas of interaction, including how platforms respond to prompts, provide educational support and account for social context. It also examines culturally responsive design and whether interactions are age-appropriate.
The authors’ observations point to a disconnect between the principles expressed in current governance frameworks and their practical application in situations involving youth well-being.
From broad principles to youth-centered practice
Many AI platforms are guided by ethical frameworks that promote user safety. However, the article notes that these frameworks do not always become youth-centered practices. A framework may state an intention to prevent harm without fully addressing the developmental, social and cultural circumstances that can shape a young person’s interaction with an AI system.
The authors call for standards specifically intended to ensure that generative AI platforms respond to those needs, along with strict adherence to the standards in practice. They also identify culturally responsive and trauma-informed design recommendations as potential ways to move AI systems toward more ethical approaches.
The article presents these recommendations as having the potential to improve AI design and reduce possible harm; it does not describe those outcomes as guaranteed.
Why the policy gap matters
For families, educators and others who encounter AI in youth settings, the discussion highlights an important distinction: the existence of a safety policy does not by itself show how a platform will respond in a specific interaction.
The perspective therefore shifts attention from whether platforms have ethical commitments to how consistently those commitments appear in prompt responses, educational support and culturally sensitive interactions. The authors argue that youth well-being should be built into these practical details rather than treated only as a general policy goal.
The article was written by Siva Mathiyazhagan, Lucia M. Nazzaro and Desmond U. Patton of SAFELab at the University of Pennsylvania. The authors declared no competing interests.
AI tools were used to assist with the preparation of this article under our editorial standards.
Leave a Reply