# What exactly does 'ChatGPT for Teens' do?

According to the New York Times, the new mode automatically restricts certain topics and adds safety guardrails when the system detects a user is under 18. OpenAI said the feature is designed to 'better protect young users' by limiting conversations on topics like self-harm, violence, and sexual content.

But here's the catch: the system relies on users to provide their age at sign-up, or on weak signals like email address. There's no biometric verification, no government ID check, and no parental consent mechanism. A 15-year-old can simply enter a different birth year and access the full, unfiltered ChatGPT.

This is not a safety solution; it's a compliance checkbox. OpenAI is not actually protecting teens—it's protecting itself from liability.

# Why is OpenAI doing this now?

The timing is no accident. The New York Times reported the announcement comes as 'safety concerns grow' among parents, educators, and lawmakers. In the US, the Children's Online Privacy Protection Act (COPPA) is being updated, and the EU's Digital Services Act is forcing platforms to assess risks for minors.

OpenAI is under pressure from regulators like the Federal Trade Commission, which has already fined companies for failing to protect children's data. By launching a teen mode, OpenAI can claim it's being proactive. But regulators are not naive—they will ask how the system verifies age, and OpenAI's answer will be 'self-declaration.' That's not an answer; it's an admission of weakness.

# Who actually benefits from this feature?

Short-term, OpenAI's PR team benefits. They get a headline that says 'we care about kids.' But the real winners are competitors like Google and Anthropic, who can now point to OpenAI's superficial approach and offer a more robust alternative.

Anthropic, for example, has already invested in age-verification partnerships and publishes detailed safety research for educational use. According to Anthropic's published materials, they are working with schools to integrate Claude into classrooms with teacher oversight and verified student accounts. That's a fundamentally different approach—one that puts the burden on institutions, not individuals.

Google's Gemini for Education also requires school-administered accounts, providing a layer of verification that OpenAI's consumer-grade teen mode lacks. In the education market, which is the most lucrative entry point for youth AI, OpenAI is now at a competitive disadvantage.

# What does this mean for the future of AI safety?

This move reveals a critical truth: AI safety for minors is not a feature, it's an infrastructure problem. The real challenge is age verification, not conversation filtering. Until companies solve that, any 'teen mode' is just a placebo.

Startups like Yoti, which specialize in age assurance, are the hidden winners here. They will likely see increased demand from AI companies desperate to meet regulatory requirements without building their own systems. The EU's upcoming AI Act mandates strict age verification for high-risk AI systems, and the market will need to comply.

OpenAI's move also sets a precedent: it's the first major AI company to launch a dedicated teen mode. This forces others to follow, but it also lowers the bar. If OpenAI can get away with self-reported age, why would anyone invest in better solutions?

# What should OpenAI do next?

First, stop pretending self-reported age is a safeguard. Second, partner with a third-party age-verification service like Yoti to actually verify users. Third, offer a separate, institution-managed version for schools that includes teacher dashboards and granular controls, similar to what Google and Anthropic already offer.

If OpenAI does not do this, they will lose the education sector entirely. Schools are risk-averse and will not adopt a product that could expose them to litigation. The New York Times reported that some school districts have already banned ChatGPT due to safety concerns; this new mode is unlikely to change their minds unless it comes with real verification.

The clock is ticking. The EU AI Act's high-risk provisions take effect in 2027, and the FTC is actively investigating children's privacy. OpenAI needs to act now, not after the next scandal.

# Predictions

1. By Q3 2027, OpenAI will announce a partnership with Yoti or a similar age-verification provider, following pressure from EU regulators.
2. By Q1 2027, at least one US state attorney general will open an investigation into ChatGPT's teen mode for failing to prevent minors from bypassing restrictions.
3. Within 18 months, Google's Gemini for Education will overtake ChatGPT in US school districts, citing verified accounts as the key differentiator.

• OpenAI's teen mode is a compliance artifact, not a real safety solution.
• Age verification is the true bottleneck, and startups like Yoti are the hidden winners.
• Google and Anthropic are already ahead in the education market with institutional accounts.
• Regulatory pressure will force OpenAI to make a real commitment within 12 months.
• The teen mode may actually increase risk by giving parents a false sense of security.