Common Sense Media questions ChatGPT for Teens’ safety safeguards
Common Sense Media says OpenAI’s ChatGPT for Teens poses an “unacceptable risk”, arguing that its safety features do not reliably protect young users. The assessment says the service may fail to alert parents about serious concerns, respond appropriately in a crisis or prevent students from getting homework answers.
OpenAI disputes the findings, saying the testing may have taken place before parental controls were fully activated and therefore may not reflect how the safeguards work. Common Sense Media says it confirmed other teen features were launched before testing, and that some linked test accounts still received no alerts even after the activation period OpenAI described. It stands by its conclusion that parental notifications are unreliable in crisis situations.
- Common Sense Media calls ChatGPT for Teens an unacceptable risk.
- Its assessment says parental alerts can fail in crisis situations.
- OpenAI disputes the testing and says it may not reflect active safeguards.
New here? Start with this
ChatGPT is an AI chatbot made by OpenAI that can answer questions and help with tasks such as writing and schoolwork. OpenAI offers a version for teenagers, with settings intended to give parents more oversight of how young people use it.
Common Sense Media is a US non-profit organisation that reviews media and technology for families. Its safety assessment raises broader questions about how well AI services can protect teenagers, support them in a crisis and fit into schoolwork.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Common Sense Media's concern centres on child protection. Their testing found that parental notifications for crisis situations were unreliable, which undermines the core safety mission parents depend upon when their teenagers use the platform. They have independently verified their testing conditions and confirmed that alerts failed even after accounting for OpenAI's activation timeline, suggesting the safeguard failures are genuine rather than methodological artefacts. For a service specifically designed for young users, consistently notifying parents of serious concerns is non-negotiable.
The case against
OpenAI's position reflects a legitimate point about product evaluation. Testing a platform before its systems are fully activated and deployed produces results that may not represent actual user experience. Parental controls and safety features are designed to work within properly configured systems; assessing them before that configuration is complete risks drawing misleading conclusions. OpenAI argues that fair assessment requires testing under realistic operational conditions, not partial deployment scenarios, to determine whether the product genuinely fails users or whether the evaluation captured an unfinished state.
Read the full article at the source →
Originally published by The Verge as “ChatGPT for Teens is an ‘unacceptable risk,’ says Common Sense Media”.