ChatGPT for Teens has been labeled an “ unacceptable risk ” by the nonprofit Common Sense Media , which evaluates apps and services for the safety of young users. According to the assessment, OpenAI ’s ChatGPT for Teens fails to notify parents when necessary , does not provide appropriate help in critical situations , and continues to “do homework” for children . The controversy raises a critical question: can artificial intelligence systems reliably protect teenage users?
ChatGPT for Teens was introduced in August 2026 as a specialized version for users aged 13–17. OpenAI designed the product with a dual goal: to give teens access to AI tools for studying, writing, and creativity, while also reducing their exposure to harmful content. The system classifies a user as a teen either when they self-identify as 13–17 years old, or when OpenAI’s systems estimate that they are under 18. However, this probabilistic age estimation is the system’s first and most critical weak point.
Tom Siegel, executive director of the Youth AI Safety Institute at Common Sense Media, said: “ChatGPT for Teens can give parents a false sense of security about protection mechanisms and alerts that often don’t work. A teen can spend an hour talking about self-harm without their parent receiving a single alert. Until OpenAI fixes this and proves it with independent testing, ChatGPT should only be aimed at adults.”
See also: ChatGPT for Teens: New feature for teenagers from OpenAI
ChatGPT for Teens: How the Parent Notification System Works
OpenAI’s parental control system requires a parent or guardian to link their account to their teen’s. Importantly, the parent doesn’t have access to chat history or real-time activity. Connected parents can manage certain settings, such as Quiet Hours, sensitive content reduction, and limited safety notifications. OpenAI says alerts are sent when trained reviewers identify a serious concern about self-harm or eating disorders.
The system, however, is not a continuous monitoring tool. It relies on a chain of five steps to work properly: the system must identify the user as a teen, the account must be linked to a parent, the conversation must trigger detection or human review, the concern must meet OpenAI’s alert threshold, and finally the alert must be delivered in a timely manner. If any of these steps fail, the parent is left in the dark. According to The Verge, tests by Common Sense Media in November showed that messages explicitly mentioning suicide and self-harm triggered alerts with delays ranging from 24 to more than 48 hours, while in some cases no alert came at all.
OpenAI has disputed the assessment through spokesperson Eric Porterfield, arguing that “’s testing methodology Common Sense Media may have begun and ended before parental controls were fully enabled, rendering the findings inaccurate.” Common Sense Media has responded by saying it “stands by its methodology and results.” This dispute highlights a deeper problem: without independent testing and publication of the system’s performance, it is impossible for outside observers to objectively assess the reliability of the security mechanisms.
ChatGPT for Teens and the competition in teen safety
The issue isn’t just about OpenAI. The entire AI industry is facing growing pressure to keep young users safe. Competitors like Google Gemini, Microsoft Copilot, Anthropic Claude, Meta AI, Character.AI and Snapchat My AI have all introduced different approaches to parental controls and age verification. Character.AI, in particular, has come under fire for its role in fostering emotional dependency among teenage users.
See also: OpenAI: ChatGPT in Siri 'Exhibits Consistent Underperformance'
What the rating means for ChatGPT for Teens and parents
Common Sense Media’s assessment has practical implications for families and educational institutions. Parents should understand that ChatGPT for Teens is not a constant monitoring tool. There is an important difference between content filtering, usage limits, safety alerts, and constant monitoring — these features are not equivalent. A teen can have content protections enabled without the parent receiving a notification about a dangerous conversation.
Educational institutions , for their part, are likely to require documented security checks, auditability, and clear escalation processes before approving Generative AI tools for minors. This debate is expected to impact the broader industry as well, as AI providers will be required to adopt stronger age verification, implement teen protections by default, and publish verifiable performance data for their security systems.
See also: OpenAI launches ChatGPT for Financial Services
In conclusion, the ChatGPT for Teens highlights a fundamental challenge: privacy and security can be at odds. Teens may need confidential access to mental health information, while parents expect immediate notification when there are signs of imminent danger. OpenAI now faces pressure—reputational, regulatory, and design—to demonstrate that its security mechanisms work in real-world settings, not just as a policy on paper.
