OpenAI launched ChatGPT for Teens on August 18, a dedicated experience for users under 18 that blocks sexual and romantic roleplay, deploys an automated age-prediction system, and gives parents Quiet Hours controls — setting a de facto standard for how AI platforms handle underage users that companion AI companies will be measured against.
The launch arrives years after teens were already using the platform, and it formalizes safeguards that OpenAI has been testing since September 2025. The company says ChatGPT is intended for users 13 and older, and its new age-prediction system will estimate a user's age based on behavior, defaulting to the teen experience whenever there is doubt.
What ChatGPT for Teens Blocks
The teen experience applies age-appropriate content protections by default that reduce exposure to several categories directly relevant to the companion AI space:
- Sexual or romantic roleplay — explicitly blocked for teen accounts
- Violent roleplay — restricted
- Graphic content and viral challenges — filtered
- Extreme beauty ideals — reduced exposure
These restrictions go beyond what standard ChatGPT accounts receive. OpenAI's Model Spec has been updated to include teen-specific protections, and the company published a Teen Safety Blueprint in November 2025 outlining its approach.
Parental Controls Without Message Access
Parents can link their account to a teen's account through an email invitation. Once linked, parents can:
- Set Quiet Hours when ChatGPT cannot be used
- Turn off voice mode
- Disable memory and, in some configurations, chat history
- Remove access to image generation
- Opt the teen's conversations out of model training
- Enable Study Mode that limits responses to learning-focused flows
Critically, OpenAI says parental controls do not let parents read or monitor a teen's actual conversations. Safety notifications are sent only in limited high-risk situations, such as potential self-harm concerns, and share only the information needed to support the teen's safety.
Age Prediction: The Regulatory Template
OpenAI's age-prediction system is the most notable policy-relevant feature. The company says it estimates age based on how people use ChatGPT, and if there is doubt, it defaults to the under-18 experience. This approach mirrors what several state laws — including California's SB 243 — are beginning to require for companion chatbots that may serve minors.
California's companion chatbot law, which took effect January 1, 2026, requires operators to take reasonable steps to prevent chatbots from encouraging minors to engage in harmful or illegal behavior. OpenAI's teen launch effectively demonstrates one compliance path: automated age gating paired with blanket content restrictions.
What This Means for Companion AI
The restrictions on sexual and romantic roleplay are the most directly relevant to AI Haven's audience. OpenAI is the most visible AI platform in the world, and its decision to explicitly ban this category for under-18 users creates a benchmark that regulators, school districts, and parents will expect from every AI service — including dedicated companion apps like Character.AI, Replika, and others.
Companion platforms that serve users under 18 now face a clear expectation: if OpenAI can detect and block sexual roleplay for teens, so can you. The age-prediction system also raises the stakes for platforms that have relied on self-reported age, since OpenAI is now proactively estimating age based on usage patterns.
The launch also extends teen protections across OpenAI's broader product line, including group chats, the ChatGPT Atlas browser, and the Sora app, signaling that age-appropriate design is becoming a company-wide policy rather than a single-product feature.