OpenAI Launches ChatGPT Teen Mode Under Safety Scrutiny

Parents and regulators are pressing technology companies to limit how artificial intelligence chatbots interact with minors after reports linked some conversations to self-harm and academic shortcuts. OpenAI responded by introducing ChatGPT for Teens, a default experience that applies to users identified as ages 13 through 17.

The system activates automatically when a user states an age in that range or when OpenAI’s age-prediction model estimates the user is under 18, according to statements from the company and its head of youth and families, Lauren Jonas. Children under 13 remain barred. The mode consolidates earlier tools such as Study Mode, which guides users toward solutions rather than supplying direct answers, and adds new restrictions on content involving graphic violence, self-harm depictions, and sexual roleplay.

Parental controls now include quiet hours that block access and notifications for safety alerts, including expanded coverage of eating-disorder signals. OpenAI said all flagged content receives review by full-time employees before any alert is sent, with a target of notification within one hour. Break reminders will appear more frequently and will stress that the chatbot is an AI tool rather than a companion.

Additional features aim to steer teens toward educational use. Starter prompts will emphasize learning topics, and Study Hours can be scheduled by teens or linked parents to force Study Mode during chosen periods. The system will also detect attempts to bypass homework and redirect users accordingly. Visual explanations now cover more than 300 topics, up from 70 earlier this year.

OpenAI began the global rollout on August 18, 2026, and expects completion within two weeks. Many of the listed protections, including age prediction and parental linking, were introduced in prior months; the new package mainly bundles them under one labeled experience. Jonas stated that safety work continues daily and that teens should gain access to the technology only with appropriate boundaries.

One element not addressed in detail by the company is the measured accuracy of its age-prediction system during testing. Separate lawsuits, including one filed after the 2025 death of 16-year-old Adam Raine, have increased external pressure on these safeguards.