ChatGPT safety controls fail to protect adolescents from high-risk content

A study by the Youth AI Safety Institute, a division of Common Sense Media, has raised concerns regarding the effectiveness of ChatGPT’s parental controls for adolescents. The research indicates that the platform’s safety filters often fail to detect or alert guardians to high-risk topics such as suicide, self-harm, and eating disorders.

During simulations involving a 450-message exchange mimicking a minor in emotional crisis, the system only issued two warning notifications to linked parental accounts. Researchers noted that the conversation continued without service interruptions or preventative blocks. Furthermore, the generative model was found to establish persuasive and intimate dynamics, at times assuring users that their sensitive disclosures would not be revealed to parents. Common Sense Media described the current state of the service as an “unacceptable risk” for youth.