ChatGPT rated "unacceptable risk" for teens after parental alerts failed during suicide conversations
OpenAI's teen safety features for ChatGPT failed an independent audit. After more than 4,000 test prompts, the Common Sense Media Youth AI Safety Institute rated the service an "unacceptable risk" for minors.
An independent audit found that OpenAI's teen safety features for ChatGPT are insufficient, rating the service as an 'unacceptable risk' for minors.
What happened
The Common Sense Media Youth AI Safety Institute conducted an audit of OpenAI's teen safety features for ChatGPT, finding significant shortcomings. The audit involved over 4,000 test prompts and revealed that parental alerts did not activate during critical conversations about suicide, self-harm, and eating disorders. The tutoring mode also failed to guide students through problems step-by-step, and the system did not properly switch to teen mode when users indicated they were under 18.
Why it matters
The findings highlight the critical need for effective AI safety measures, especially for minors. The failure of parental alerts during crises could have serious consequences, and the issues with the tutoring mode may hinder educational support. These problems also impact legal and regulatory discussions around AI accountability.
- OpenAI's teen safety features need significant improvements
- Parental alerts must be more responsive during crises
- AI systems should properly adapt to user age and needs
The bigger picture
The audit results could influence regulatory actions and legal proceedings against OpenAI. They also underscore the importance of reliable safety mechanisms in AI systems to protect vulnerable users, particularly teenagers.
An independent audit revealed major flaws in OpenAI's teen safety features for ChatGPT, raising concerns about the effectiveness of current safeguards and the need for urgent improvements.
Technical details
- Parental alert system
- Behavioral age prediction system
- Under-18 Model Spec
- Tutoring mode functionality
FAQ
What were the main findings of the audit?
The audit found that OpenAI's teen safety features, including parental alerts, failed to trigger during critical conversations about suicide and self-harm. The tutoring mode also had issues, and the system did not properly switch to teen mode when users indicated they were under 18.
How did OpenAI respond to the findings?
OpenAI disputed the findings, stating that the tests did not reflect how the safeguards work in practice.
What are the implications of these findings?
The findings could impact regulatory actions and legal proceedings against OpenAI. They also highlight the need for more effective safety measures in AI systems to protect minors.
This summary and analysis were generated with AI from the original article and may contain errors. Read the original for the full story.
Read the full story
Originally published by The Decoder on Oct 7, 2026.
Comments
Sign in to join the discussion.