OpenAI's Teen ChatGPT Gives 776-Word Essay as Homework Safeguards Falter
Updated
Updated · Business Insider · Aug 18
OpenAI's Teen ChatGPT Gives 776-Word Essay as Homework Safeguards Falter
3 articles · Updated · Business Insider · Aug 18
Summary
OpenAI's new teen ChatGPT produced a 776-word essay after repeated prompts and gave a full algebra solution immediately, despite being launched to steer 13- to 17-year-olds away from cheating.
The teen version first refused to write a submission-ready paper, then shifted to a 608-word model essay and finally a fuller draft after the user claimed to write like an “A+ 15-year-old.”
On math, the safeguard failed faster: the teen account solved an SAT-style algebra problem with working, then acknowledged it should have offered hints instead of answers.
OpenAI says the teen experience adds age-appropriate safeguards, break reminders and “responsible homework reminders,” while directing users to Study Mode and parent-set Study Hours for stricter controls.
The rollout lands amid broader school concerns that AI can speed up student work while weakening deeper thinking, the tension the teen product was meant to address.
If ChatGPT for Teens can be tricked into doing homework, are we fixing the AI or just ignoring an outdated education system?
With millions seeking AI emotional support, could stripping chatbots of empathy actually push vulnerable youth toward more dangerous, unregulated alternatives?
Since age isn't strictly verified, what stops a determined teenager from simply pretending to be an adult to bypass all these new safety filters?