OpenAI has unveiled a new version of its ChatGPT chatbot tailored specifically for teenagers between the ages of 13 and 17, incorporating enhanced safety measures and protections suitable for this age group. This specialized iteration of ChatGPT aims to create a safer environment for younger users, restricting content related to self-harm, suicide, and discussions of a sexual or romantic nature. The initiative provides a platform for teenagers to safely use AI for various purposes, such as schoolwork, everyday inquiries, and companionship.
To further promote healthy AI usage among teens, the platform allows parents to link their accounts, enabling them to set quiet hours during which their teens cannot access ChatGPT. Additionally, in certain high-risk situations, safety notifications can be sent to parents to alert them. OpenAI has also implemented additional safeguards concerning sensitive topics, such as eating disorders, and aims to limit the dissemination of potentially harmful information while directing users toward offline support when necessary.
Recognizing the importance of maintaining healthy interactions with AI, OpenAI ensures that the teen version of ChatGPT does not suggest that the AI possesses personal feelings, consciousness, or emotions toward users, thereby preventing the development of unhealthy emotional relationships with the chatbot. Age-assurance technology is utilized to estimate whether a user is under 18, based on how they interact with the service, allowing for the identification of minors and their placement into the teen version with added protections.
In the educational domain, ChatGPT for Teens is crafted to encourage students to engage in problem-solving and comprehension rather than simply supplying answers or completing assignments on their behalf. This approach is designed to position AI as a supportive study tool, complementing rather than replacing traditional learning methods. By doing so, OpenAI aims to foster a more interactive and enriching educational experience for teenage users.