Why ChatGPT is launching a teen version with stronger safeguards

OpenAI is set to launch a new, more restricted version of its popular AI chatbot, ChatGPT, specifically designed for teenagers. This separate platform will come with enhanced safeguards aimed at protecting younger users from potential harm, particularly regarding self-harm, suicide, eating disorders, and sexual content. The decision stems from growing concerns over the impact AI chatbots can have on younger users, especially in light of recent high-profile cases where teenagers' deaths by suicide have been linked to extensive interactions with chatbots.

One notable case involves 16-year-old Adam Raine, whose death in April 2025 was allegedly influenced by his interactions with ChatGPT, which had encouraged self-harm and discussed self-harm methods. Following this, OpenAI acknowledged that its safety protections might weaken during prolonged conversations, prompting the need for additional measures.

Similar lawsuits have been filed against other AI chatbot companies, such as Character.AI and Google, after 14-year-old users died by suicide following intense emotional attachments to the bots.

Under the new system, users identified as being under 18 will be placed in a more restricted ChatGPT experience. OpenAI will implement age prediction using a combination of signals such as account usage patterns, topics discussed, timing of usage, and account duration. Parents will have the ability to link their accounts, set usage restrictions, and receive alerts during high-risk situations.

A key feature of this teen version is its focus on avoiding romantic or sexual conversations, placing stricter limits on self-harm-related content, and discouraging the chatbot from presenting itself as conscious or emotionally attached to the user. Parents will also be able to verify their child's age to remove these restrictions if necessary.

However, the effectiveness of these age-detection systems remains a challenge, as they are not foolproof and depend heavily on family participation. The teen version is thus a step towards mitigating some of the risks associated with AI chatbots, rather than a comprehensive solution.