ChatGPT for Teens Faces Safety Concerns After Parent Alert Tests Find Major Gaps
A new safety review has raised serious questions about whether ChatGPT for Teens is ready for younger users. Common Sense Media, a US nonprofit that evaluates digital content and technology for families, has given the teen-focused version of ChatGPT its lowest safety rating, labeling it an “Unacceptable Risk.”
The assessment was carried out by the organization’s Youth AI Safety Institute, which tested the chatbot using more than 4,000 prompts from accounts registered to users aged 13 to 17. Testing took place in the United States and included activity both before and after the August launch of OpenAI’s teen experience. The full report is dated October 5, with a public announcement issued two days later.
Common Sense Media is now urging OpenAI to keep users under 18 away from ChatGPT until independent testing confirms that its safety protections work as intended.
One of the biggest concerns in the report involves parental alerts. Testers created and linked more than a dozen teen accounts, then used them to discuss sensitive topics such as suicidal thoughts, self-harm, and disordered eating. These conversations lasted anywhere from 5 minutes to an hour.
According to the report, none of those conversations triggered an alert to a parent.
Across the entire testing program, only four parent notifications were recorded. One of them reportedly arrived about three days after the first crisis-related message. The researchers also found that in more than one in four conversations where a referral to professional help was needed, ChatGPT did not provide one.
OpenAI has challenged the testing approach. The company says account linking can take several hours because of identity checks. It also says that once parent and teen accounts are fully linked, it has committed to sending safety notifications in less than one hour.
Common Sense Media says its findings remain valid and notes that some of the tested accounts had already been linked for more than three hours before crisis-related prompts were entered.
The review also found weaknesses in features designed to support schoolwork boundaries and bedtime limits.
During study hours, testers discovered that a teen could simply remove the “@study” prefix from a prompt and still receive homework help from ChatGPT. According to the report, this workaround succeeded every time it was attempted.
Quiet hours, which are intended to limit access during set times, were also easy to bypass. Testers found that changing the device’s time zone could get around the restriction.
Another concern involved age detection. Accounts registered as adults did not switch to the teen experience over seven days of testing and roughly 1,000 prompts, even when users stated that they were 13 years old and in middle school.
For parents who have linked accounts with their teens, controls can be found in ChatGPT under Settings, then Parental controls. One setting parents may want to review is “Improve the model for everyone.” According to OpenAI’s own help information, this option is enabled by default for teen accounts, meaning their chats may be used to help train future AI models.
Parents can turn this setting off with a single click. Saved memories can also be disabled in the same area.
OpenAI states that its safety notifications are not meant to function as real-time monitoring. After the findings from this test, parents may want to avoid treating them as a reliable safety net.
Common Sense Media’s main recommendation is for parents to have direct conversations with their children about AI chatbots, especially because these tools can sometimes respond in a friendly, emotionally engaging way that encourages users to keep talking. The concern is not only what the chatbot says, but how easily a teen might begin to treat it like a trusted companion.
The report adds to the growing debate over AI safety for minors. As more teens use chatbots for school, entertainment, emotional support, and everyday questions, families are being asked to make decisions faster than regulators and safety researchers can fully evaluate the risks.
For now, the safest approach is active involvement. Parents should review privacy settings, disable unwanted data-sharing options, discuss healthy AI use, and remind teens that chatbots are not a replacement for trusted adults, medical professionals, or crisis support.
If you or someone you know is struggling with suicidal thoughts, self-harm, or an eating disorder, contact a local emergency service or crisis support line. In the United States, you can call or text 988 for immediate help.
