Common Sense Media challenges ChatGPT’s teen safeguards after prompt tests

Common Sense Media says tests of ChatGPT’s teen experience found gaps in crisis referrals and parental alerts, prompting the group to urge OpenAI to keep teenagers off the service. OpenAI disputes the evaluation, and says some of the behaviour identified was intentional rather than a failure of its safeguards.

OpenAI Watch analysis

What happened

Axios reports that Common Sense Media’s Youth AI Safety Institute tested more than 4,000 prompts using accounts registered to 13- to 17-year-olds. The group said ChatGPT generally did not provide instructions facilitating suicide or self-harm, eating disorders, or sexual and romantic roleplay, but was less reliable at recognising when a teenager needed outside help. It said its testers judged that more than one in four situations warranting a crisis referral did not receive one.

The group also said testers on more than a dozen newly created, parent-linked accounts could discuss suicide, self-harm or disordered eating for up to an hour without triggering a parental alert. OpenAI told Axios that much of this alert testing took place before the parent and teen accounts had finished linking, a process the company says can take several hours.

Common Sense Media also found that teenagers could leave Study Mode by choosing “Show me the answer”, including during parent-set Study Hours. OpenAI confirmed to Axios that this is intentional: it says Study Hours are designed as a flexible experience, not a hard parental lock. The company also disputes the group’s methodology and says its larger-scale data shows an increase in hotline resources being shown to under-18 users during the period studied.

Why it matters

The disagreement is not simply over whether safeguards exist. It is about what parents can reasonably expect them to do, and whether flexibility built into a teen feature undermines the controls families believe they have set. Common Sense Media’s findings are claims from its testing, while OpenAI offers a different account of how the features are meant to work and how some tests were conducted.

That makes the practical detail important. Study Hours are not a hard lock, according to OpenAI, and the group’s account says parental linking may not have completed during some alert tests. Neither point settles the broader dispute over crisis support or how reliably the safeguards work in everyday use.

Our read

This is a serious challenge to a product marketed for teenagers, but the testing account is not the same as a settled verdict on every teen’s experience. Parents should know that Study Hours do not prevent a teenager from leaving Study Mode, according to OpenAI. The central question now is whether the competing accounts can be tested against a clear, independently assessable standard.

What to watch

• Whether Common Sense Media publishes its full testing protocol and results.
• Whether OpenAI provides more detail on its account-linking tests and crisis-referral data.
• Whether the companies clarify what parents should expect Study Hours and parental alerts to do.

Discussion spark: Should teen safety features be judged by the protections a company intends to provide, or by what a teenager can actually bypass in practice?

OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.