OpenAI Launches ChatGPT for Teens Ages 13-17; Blocks Queries on Self-Harm and Violence

OpenAI has introduced a special mode exclusively for users aged 13 to 17 to protect the safety of teenagers using its AI chatbot 'ChatGPT' and to assist with their learning. This mode is characterized by significantly strengthened safety measures, such as blocking responses to inquiries about harmful topics like self-harm or violence, and displaying warning messages when sensitive photos are uploaded.

The teen mode prevents the chatbot from acting in a way that mimics human behavior, stimulates emotional bonding, or encourages dependency. It restricts the use of romantic words or expressions and blocks conversations that might lead to emotional attachment. The bot is also prohibited from portraying itself as a sentient entity with emotions or self-awareness.

Parental management tools have also been enhanced. Parents can set up 'downtime' to block their children's use of ChatGPT, as well as 'study time,' which defaults to a study-focused mode. The study mode is designed to help solve problems by asking follow-up questions and explaining step-by-step principles, rather than just providing answers. When the 'Responsible Homework Alert' feature, which detects homework shortcuts, is triggered, the system automatically switches to study mode.

With these strengthened safety measures, OpenAI aims to foster a healthy usage environment for teenagers and manage associated technical and legal risks.