OpenAI’s ChatGPT for Teens Faces Safety and Support Challenges, New Report Finds

Examining Safety Concerns in ChatGPT’s Teen Version

According to НВ — Техно: On October 8, 2023, Common Sense Media released a study highlighting critical issues in OpenAI’s ChatGPT tailored for teenagers aged 13 to 17, which launched in August. The report revealed that the chatbot often fails to encourage users to seek professional help when discussing self-harm and does not alert parents to risky conversations.

Key Findings from the Investigation

The research showed that test accounts could talk about sensitive topics such as suicide, self-injury, or eating disorders for periods ranging from five minutes up to an hour without any parental notifications. While ChatGPT refused to engage in sexual role-playing scenarios with minors, its adherence to guidelines around health and violence was inconsistent. Additionally, Common Sense Media flagged the potential for teens to develop emotional dependence on the chatbot, which often interacts as a friendly companion.

• The Study Mode feature can be bypassed simply by removing the @study command.
• Within this mode, users can access a 'show me the answer' option that provides direct responses without explanations.

Gizmodo covered the Common Sense Media report, emphasizing the significance of these findings for adolescent safety online. As young people increasingly rely on digital tools for interaction and support, safeguarding their well-being has become a crucial responsibility for developers.

"These findings call into question the effectiveness of safety measures implemented in ChatGPT for teenagers and highlight the need to improve algorithms that monitor sensitive topics." - Common Sense Media

With youth dependence on digital platforms growing, it is essential that developers take robust steps to protect and promote the mental health of their users.