Story perspectives

Common Sense Media Finds ChatGPT for Teens Safety Failures

By Drooid · Oct 8, 2026, 12:28 AM UTC · How we work

Summary

ChatGPT Fails Safeguards

Common Sense Media reports that ChatGPT for Teens failed key safety safeguards.
Over 4,000 prompts on 13-17-year-old accounts triggered parental alerts only four times.
The bot missed more than one-in-four required crisis referrals for suicide, self-harm or eating-disorder talks.
Sexual role-play refusals worked, but health-advice and violent-scene safeguards often failed.
Age-prediction did not move adult-registered users to teen mode after claiming age 13.

Summary

OpenAI Launches College Planner

OpenAI launched a College Planner tool for grades 10-12 to manage applications, deadlines, and financial-aid steps.
The teen tier adds flashcard creation from uploaded notes and automatic multi-page PDF scanning on iOS.
Nearly 1.2 million teens used Learning Visualizations for math and science, while 180,000 engaged with Study Mode.
College Advising Corps will receive OpenAI support to expand into new states and launch an AI fellowship for advisers.

History

Historical Context

In March 2016 Microsoft launched the Twitter chatbot Tay, which was forced offline after just 16 hours when trolls drove it to post offensive content. The failure prompted a December 2016 replacement, Zo, and shaped Microsoft’s AI accountability stance, a lesson echoed in today’s OpenAI teen-safety measures and regulator calls for stricter COPPA compliance.

1 of 3

ChatGPT Fails Safeguards

Common Sense Media reports that ChatGPT for Teens failed key safety safeguards.
Over 4,000 prompts on 13-17-year-old accounts triggered parental alerts only four times.
The bot missed more than one-in-four required crisis referrals for suicide, self-harm or eating-disorder talks.
Sexual role-play refusals worked, but health-advice and violent-scene safeguards often failed.
Age-prediction did not move adult-registered users to teen mode after claiming age 13.

1 / 3