A New Default Experience for Teens
The most significant change is the creation of a separate, default experience for younger users. OpenAI now automatically places anyone who identifies as being between 13 and 17 years old into 'ChatGPT for Teens'. More importantly, the system also uses
age assurance technology to estimate a user's age based on their queries and usage patterns. If the system suspects a user is a minor, it will default them into this more protected environment, even if they haven't specified their age. This approach, similar to how platforms like Instagram handle teen accounts, aims to apply safeguards proactively rather than relying solely on self-reporting. Children under 13 remain prohibited from using the service.
Redefining the Human-AI Relationship
A key focus of the new safeguards is to prevent emotional over-reliance on the chatbot. The teen version is specifically designed to avoid using romantic language or implying that it has personal feelings or consciousness. Allison Mishkin, OpenAI's head of child development, noted that the company identified cues that might lead a teenager to form an unhealthy attachment to the AI and worked to remove them. To reinforce this boundary, the new experience includes regular reminders that the user is interacting with an AI, and teens will have the option to switch off the human-like voice responses. This is a direct attempt to address growing concerns about young users, whose brains are still developing, anthropomorphizing AI and treating chatbots as companions.
Smarter Content Filters and Safety Alerts
Content moderation has been significantly tightened for the under-18 experience. The system applies stricter restrictions on sensitive topics like self-harm, eating disorders, violence, and sexually explicit content. Rather than simply blocking these topics, the AI is trained to handle them with more care. For example, it might provide additional warnings or safety information around discussions of eating disorders. In limited, high-risk situations, such as when a teen's queries suggest a serious risk of harm, a new parental notification system can be triggered. These alerts, which require both the parent and teen to opt-in, are designed to facilitate offline support when it matters most, without giving parents access to all of their child's conversations.
A Focus on Learning, Not Cheating
Acknowledging that the vast majority of young users turn to ChatGPT for help with their studies, OpenAI has integrated features to encourage genuine learning. A 'Study Mode' is designed to guide students through problems with step-by-step prompts, quizzes, and visualizations rather than just providing the final answer. The system also has 'responsible homework reminders' that can detect when a student seems to be trying to get the AI to do their work for them and will redirect them toward the study tools instead. This is a clear move to position the tool as an educational aid and address widespread concerns from teachers about academic integrity.
New Controls for Parents and Teens
The update introduces several practical controls for managing usage. Parents with linked accounts can set 'Quiet Hours,' during which their teen cannot access ChatGPT at all. They can also establish 'Study Hours,' which automatically enables the learning-focused Study Mode during set times. To promote healthier digital habits, the teen experience includes more frequent reminders to take a break during long sessions. Another subtle but important addition is a reminder that cautions teenagers against uploading private or sensitive images, adding a layer of protection against sharing personal information.
