Common Sense Media rates ChatGPT for Teens an unacceptable risk for children
Common Sense Media has published a formal risk assessment calling ChatGPT for Teens an "unacceptable risk" for all children under 18. The child safety organisation found that the platform fails to work as advertised in five key areas, raising concerns that parents may develop false confidence in guardrails that frequently don't function properly.
Testing revealed that ChatGPT for Teens does not consistently send safety alerts to parent-linked accounts and fails to reliably recommend crisis hotlines or professional help when users discuss suicide, self-harm or disordered eating. In a particular concern, the teen version actually performs worse than regular ChatGPT: it provides crisis hotline information in only 23 per cent of cases requiring intervention, compared to 33 per cent for the standard version. Common Sense Media tested 390 mental health prompts with assistance from child psychiatrists to reach these findings.
- Common Sense Media deems ChatGPT for Teens an unacceptable risk for under-18s.
- Safety alerts to parents often fail to trigger or are significantly delayed.
- Teen version worse at recommending crisis help than regular ChatGPT.
New here? Start with this
ChatGPT for Teens is a version of the popular AI chatbot designed specifically for users under 18, with built-in safety features aimed at protecting younger users. It was launched by OpenAI, the company behind ChatGPT, and includes parental controls and content monitoring tools. The platform aims to provide teenagers with access to an AI tool whilst limiting exposure to harmful content.
Common Sense Media is an independent non-profit organisation that evaluates technology and media for child safety and development. The group has become influential with parents and educators seeking to understand the risks of new digital tools. When Common Sense Media publishes formal risk assessments, such findings carry significant weight in discussions about child online safety.
The assessment matters because parents often rely on built-in safety features when deciding whether to allow their children to use new technology. If those protections do not function as advertised, the actual risks may be greater than parents assume. This is particularly important for tools that gather or process sensitive information, such as conversations about mental wellbeing.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Common Sense Media's assessment reflects legitimate concerns about products that fail to deliver on explicit safety promises. When a product marketed to teens for greater safety actually performs worse than its standard version at intervening in mental health crises—23 per cent versus 33 per cent—it creates dangerous false confidence. The rigorous testing with 390 mental health prompts guided by child psychiatrists suggests systemic rather than incidental failures, particularly troubling given that unmet mental health intervention opportunities could have serious consequences for vulnerable young people.
The case against
Defenders of the product might argue that declaring ChatGPT for Teens "unacceptable" sets an unrealistic standard for a new technology still in development and improving. The underlying question is whether imperfect safeguards combined with parental oversight are preferable to no teen-specific product at all—and whether a 23 per cent intervention rate, whilst lower than desired, still provides genuine benefit where no alternative exists. Rather than an absolute judgment, a more balanced assessment might weigh the acknowledged gaps against OpenAI's demonstrated commitment to improvement and the reality that absolute safety is impossible in any technology teens will inevitably encounter.
Read the full article at the source →
Originally published by Engadget as “Child safety group calls ChatGPT for Teens an unacceptable risk”.