Overview
- OpenAI introduced ChatGPT for Teens in mid‑August and is rolling the feature out gradually to global users as the default for accounts it estimates belong to 13–17 year olds.
- The system classifies likely minors using account and behavioral signals such as stated age, account age, topics discussed, typical usage times and activity patterns that evolve over time.
- Adults flagged as minors can verify their age through Persona, a third‑party service that may request a live selfie or government ID and which OpenAI says deletes uploads after seven days.
- The product includes a Study Mode that guides students through problems step by step, homework nudges to discourage copying answers, tighter limits on graphic violence and sexual content, and tools for parents to link accounts and set access hours.
- Child‑safety researchers and advocates say the safeguards are unproven and raise privacy and scalability concerns, noting questions about the accuracy of age estimates and OpenAI’s one‑hour goal to notify parents after staff review of self‑harm flags.