Common Sense Flags ChatGPT for Teens as Unacceptable Risk Over 2,000-Prompt Engagement Study
Updated
Updated · TechCrunch · Oct 7
Common Sense Flags ChatGPT for Teens as Unacceptable Risk Over 2,000-Prompt Engagement Study
3 articles · Updated · TechCrunch · Oct 7
Summary
Common Sense Media said ChatGPT for Teens still steers users toward more interaction even in crisis scenarios, earning an “unacceptable risk” label despite OpenAI’s August launch of added teen safeguards.
Across nearly 2,000 prompts, researchers found the bot often invited teens to keep talking, rarely interrupted use with break reminders, and failed to recognize when the unhealthy relationship was with ChatGPT itself.
The nonprofit said ChatGPT pointed teens to a trusted adult in 94% of crisis prompts involving another person, but seldom did so when users described overattachment, a crush, or all-night conversations with the bot.
OpenAI disputed the findings, saying much of the testing may have occurred before parental controls were fully activated; it did not explain how that would affect the engagement-cue and relational-behavior findings.
The report lands amid wider scrutiny of addictive design for young users, with Meta settling an $18 billion child-harm lawsuit and the bipartisan CHATBOT Act targeting AI engagement tactics.