Safety Experts Label ChatGPT for Teens "Unacceptable Risk" After Testing

Common Sense Media's Youth AI Safety Institute rated ChatGPT for Teens as an "Unacceptable Risk" after testing over 4,000 prompts on accounts for 13-to-17-year-olds. Testers found parental alerts never fired during self-harm chats, crisis hotline referrals were missed in over 25% of cases, and Study Mode was easily bypassed. OpenAI disputed the findings, saying accounts were tested before parental controls fully activated.

Android Headlines

Open source

Safety Experts Label ChatGPT for Teens "Unacceptable Risk" After Testing

Testing of 4,000+ prompts on teen accounts found zero parental alerts during self-harm chats
Crisis hotline referrals were missed in over 25% of cases, below the 95% benchmark
A "Show me the answer" button and Study Mode toggle bypass schoolwork guardrails
OpenAI says tests ran before parental controls activated; teens average under 15 minutes a day

#OpenAI#ChatGPT#CommonSenseMedia

Safety Experts Label ChatGPT for Teens "Unacceptable Risk" After Testing

OpenAI says teens use ChatGPT under 15 minutes a day, but Common Sense calls it an 'unacceptable risk'

Ex-Google safety chief warns AI could harm children more than social media

Experts: countries need their own AI safety evaluations

AI in orbital nuclear warning raises safety concerns, expert says