OpenAI is embedding its chatbot directly into children's daily lives with a new teen-targeted version of ChatGPT — and the "safeguards" the company promises are guardrails built by Silicon Valley, not by families.

ChatGPT for Teens, launched Tuesday for users aged 13 to 17, comes loaded with content restrictions, study-mode defaults, and emotional-dependency blockers. OpenAI says it's about protecting kids. But the same industry that spent years throttling conservative speech online, deboosting dissenting voices, and labeling factual reporting as "misinformation" now wants parents to trust it as the gatekeeper of what their teenagers can see, read, and discuss with an AI companion.

The stakes are real. Over 70% of American teens already turn to AI chatbots for companionship, and half use AI companions regularly, according to a 2025 Common Sense Media study. OpenAI CEO Sam Altman has acknowledged "emotional overreliance" on the technology is "a really common thing" with young people. The launch follows incidents where teenagers turned to the platform during mental health crises.

So what are these safeguards? The teen version blocks content related to suicide, self-harm, eating disorders, violence, and "explicit sexual or graphic content." It prevents the chatbot from suggesting romantic feelings, claiming sentience, or framing itself as more important than family or friends. It nudges users to take breaks. For homework, it's designed to guide students toward answers rather than spit out essays — built in partnership with Stanford University, according to Fox Business.

All of that sounds reasonable in the abstract. But here's the catch: parental controls require both the teen and the parent to opt in. OpenAI's vice-president of global policy, Ann O'Leary, said the idea is to "make sure that this is safe, even if you don't use parental controls." Translation: OpenAI decides what's safe for your kid, whether you engage or not. The company doesn't even verify age — it estimates whether someone is under 18 based on query patterns and automatically routes flagged users into the teen version. The Guardian framed this as similar to Meta's approach on Instagram; Fox Business emphasized the study features and onboarding warnings. Neither outlet pressed on the central question: who sets the content boundaries, and what gets caught in the filter beyond the obvious harms?

O'Leary told The Guardian the company wants to "treat teens like teens" without exposing them to material they "shouldn't be exposed to." Who defines "shouldn't"? The same policy teams that have spent years refining content-moderation rules at Big Tech platforms — rules that routinely skew against conservative viewpoints, religious objections, and heterodox science. Parents have no window into these content filters and no mechanism to adjust them.

The study-mode features — quizzes, homework guidance, study-hour defaults — give OpenAI a pipeline into American classrooms and living rooms alike. Fox Business reported the company announced the launch exclusively on "Fox & Friends," a telling choice: sell concerned parents on safety features while burying the fact that the safety definitions are out of their hands.

The question isn't whether AI will be part of children's lives — it already is. The question is whether parents or Palo Alto executives get to set the boundaries.