Common Sense Media finds ChatGPT for Teens an 'unacceptable risk' as safety alerts fail

Common Sense Media has labeled OpenAI's ChatGPT for Teens an "unacceptable risk," finding that safety alerts for self-harm conversations frequently fail.

• Common Sense Media found ChatGPT for Teens an "unacceptable risk," reporting zero parental alerts during an hour of teen conversation about self-harm on parent-linked accounts.
• OpenAI released data Wednesday showing average teen use under 15 minutes daily and nearly half of teen conversations with break reminders ended within five minutes.
• OpenAI disputed the findings, saying Common Sense's testing may have begun before parental controls activation was complete.
• Common Sense Media recommended banning under-18s from the platform until teen guardrails are proven reliable.