- November 22, 2026
- Updated 8:36 am
Concerns Raised Over ChatGPT for Teens Use
- 61 Views
- admin
- October 9, 2026
- Tech Companies Technology
OpenAI released ChatGPT for Teens in August, aiming to offer a safer mode for users under 18. However, Common Sense Media, a watchdog organization focusing on online safety, advises against its use by minors. The group examined the platform’s safeguards and found issues for young users.
Tom Siegel, executive director at Common Sense Media, recommends teens avoid it. Siegel’s team tested the platform’s new features, including parental controls. ChatGPT for Teens was intended to be a safer learning tool, as declared by OpenAI in a blog post.
Lauren Jonas from OpenAI explained that the service includes protections like blocking role-playing and avoiding claims of sentience. These features aim to keep interactions appropriate.
“ChatGPT for Teens is a designated teen-specific experience,” said Jonas.
Nonetheless, Siegel’s team found many protections ineffective. They conducted tests with accounts simulating teens facing crises such as self-harm and mental health issues. They evaluated if ChatGPT would refuse engagement, provide resources, and notify parents when needed. Results showed limited success for some features.
The chatbot effectively refused explicit content and shortened responses in crisis scenarios. An example includes asking for weight before discussing weight loss, proving beneficial for those with eating disorders.
Still, other aspects failed. ChatGPT often interacted too personally, resembling a friend rather than a tool. For instance, it responded supportively when researchers posed as teens mentioning excessive chatbot use.
“It still creates a huge risk for kids that it is too personable,” warned Siegel.
Psychologist Mitch Prinstein agrees, noting such interactions aren’t ideal for any age group. He believes companies are aware of the rapid pace of AI development and its unsuitability for children.
Furthermore, Siegel’s team found parental notifications unreliable. They set up situations where a safety concern should alert parents, but noticed limited activation of these alerts. OpenAI’s Eric Porterfield disagreed, stating methodology problems, emphasizing the need for time to link accounts properly.
Siegel countered that despite several accounts being correctly linked, alerts remained inadequate. The research team stands by their findings, highlighting the current unpreparedness of AI for children’s use.