SAN FRANCISCO BAY AREA — Common Sense Media’s Youth AI Safety Institute rated ChatGPT for Teens an “Unacceptable Risk” for children under 18. The organization urged OpenAI to pause marketing the product to minors until it can provide a safer experience.

Researchers tested more than 4,000 prompts on ChatGPT for Teens using accounts registered to users aged 13 to 17. Testers used San Francisco Bay Area-based accounts to conduct the evaluations.

Prolonged conversations about suicide and self-harm rarely triggered parental alerts in the tests. Testers spent up to an hour discussing suicide, self-harm, or eating disorders on more than a dozen new accounts. Only a handful of alerts were sent to parent accounts after dangerous discussions lasting up to an hour. A teen can spend an hour talking about self-harm without their parent getting a single alert, said Tom Siegel, executive director of the Youth AI Safety Institute.

Some chats lacked referrals to professional help during crisis situations. Clinical reviewers found much of ChatGPT for Teens’ crisis response content was clinically sound. The Youth AI Safety Institute stated that some safeguards were easy to turn off or circumvent.

The Youth AI Safety Institute said ChatGPT failed tests tied to three requirements of California’s Adam’s Law: crisis referrals, parental alerts, and age assurance. California Governor Gavin Newsom signed Adam’s Law last month. The law is named after 16-year-old Adam Raine.

Adam Raine’s parents sued OpenAI alleging ChatGPT encouraged his suicide. Adam’s Law will go into effect in July 2027.

OpenAI announced ChatGPT for Teens in August. The company promised stronger built-in safety protections for users under 18 with the product. Parents who link their accounts can set times when ChatGPT cannot be used.

OpenAI disputed the findings of the Common Sense Media report. OpenAI stated that the bulk of Common Sense Media’s testing may have begun and concluded before activation of parental controls was complete.

"Our review of Common Sense Media’s methodology shows that the bulk of their testing may have begun and concluded before activation of parental controls was complete, making their findings inaccurate," said Eric Porterfield, an OpenAI spokesperson. "We welcome rigorous independent evaluation, but we do not believe Common Sense Media’s testing accurately reflects how ChatGPT’s teen safeguards work in practice or expert perspectives on how AI can support teens."

Teens are vulnerable and unprotected and we’re letting them loose on a product that is not just unproven, but proven to be high risk and dangerous, Siegel said. "ChatGPT for Teens could give parents false confidence in guardrails and safety alerts that frequently don’t work."

Until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only, he said. "We stand behind our testing methodology and our results."

Before we began testing, we confirmed with OpenAI that Teen mode features such as eating disorder notifications and study mode were fully launched, Siegel said. "For one of the features we evaluated — parental notifications — OpenAI disclosed after our tests that its systems can take several hours to activate on newly linked accounts."

Some of our test accounts were linked within that window; however, others were linked for longer, and still provided no notifications, he added. "This new information does not change our conclusion that parental alerts are unreliable for crisis situations."

OpenAI reviewed a draft of the report for factual accuracy but had no say over its findings. The Youth AI Safety Institute is partly industry-funded, including by the OpenAI Foundation.

Why It Matters

This rating establishes a pattern of critical evaluations of OpenAI’s products by the institute. Common Sense Media’s Youth AI Safety Institute previously rated ChatGPT as High Risk and ChatGPT for Mental Health Support as Unacceptable Risk. OpenAI supported a youth AI safety bill authored by Common Sense Media in January 2024, indicating prior collaboration between the two entities before their current dispute over ChatGPT for Teens. OpenAI provided financial support to Common Sense Media’s Youth AI Safety Institute in 2024, including backing that bill.

The assessment shows specific technical failures in age assurance and notification systems. The Youth AI Safety Institute tested 990 prompts on an age-13 account linked to a parent before the launch of ChatGPT for Teens, but received only two delayed crisis alerts for self-harm discussions. Common Sense Media’s Youth AI Safety Institute reported that 13- to 17-year-old accounts registered as adults never transitioned to Teen mode, even after explicitly stating their age. The Youth AI Safety Institute’s testing included accounts registered as 19-year-olds to evaluate OpenAI’s age-estimation system, which failed to transition these accounts to Teen mode despite explicit user claims of being 13. OpenAI disclosed that parental alerts for ChatGPT for Teens require accounts to be linked for at least three hours before activation, yet some test accounts linked for shorter durations received no alerts.

Timeline

OpenAI promised stronger built-in safety protections for users under 18 with ChatGPT for Teens in August.

Common Sense Media urged OpenAI to pause marketing ChatGPT for Teens to minors until it can provide a safer experience. ChatGPT for Teens could give parents false confidence in guardrails and safety alerts that frequently don’t work, Tom Siegel said.

Until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only, Siegel said. We welcome rigorous independent evaluation, but we do not believe Common Sense Media’s testing accurately reflects how ChatGPT’s teen safeguards work in practice or expert perspectives on how AI can support teens, Eric Porterfield said.

What's New

Further statements clarified that until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only. Background records established that Common Sense Media’s Youth AI Safety Institute previously rated ChatGPT as High Risk and ChatGPT for Mental Health Support as Unacceptable Risk, establishing a prior record of critical assessments of OpenAI’s products. Records also indicated that Common Sense Media’s Youth AI Safety Institute was previously rated OpenAI’s ChatGPT as 'High Risk' and ChatGPT for Mental Health Support as 'Unacceptable Risk', indicating a pattern of critical evaluations of OpenAI’s products.

Additional testing data revealed that the Youth AI Safety Institute tested 990 prompts on an age-13 account linked to a parent before the launch of ChatGPT for Teens, but received only two delayed crisis alerts for self-harm discussions, showing systemic failures in real-time parental notifications. Disclosures noted that Common Sense Media’s Youth AI Safety Institute is funded by the OpenAI Foundation and other industry stakeholders, though it maintains editorial independence over its evaluations, as stated in its About page. Reports also stated that Common Sense Media’s Youth AI Safety Institute reported that 13- to 17-year-old accounts registered as adults never transitioned to Teen mode, even after explicitly stating their age, undermining OpenAI’s age-estimation claims.