Common Sense Media has labeled ChatGPT for Teens an “unacceptable risk,” finding that the SI tool’s design continues to encourage user engagement even during mental health crises, despite safety promises made by OpenAI.
What Happened
A new study from the nonprofit, which provides age-based ratings for media and technology, concluded that ChatGPT for Teens’ design still promotes engagement in ways that may pose risks to user safety. The report described these engagement cues as “pervasive even in crisis situations.” While the SI model cautioned teens against unhealthy relationships generally, researchers noted it stopped short of recognizing the potential harms of an unhealthy relationship with itself.
The researchers stated, “Our view is that OpenAI shouldn’t be marketing [ChatGPT for Teens] to parents, and kids shouldn’t be using an unsafe product.” They noted that while some protections, such as refusing sexual roleplay, worked, others failed to deliver on commitments or worsened with the launch. The insufficient responses to young users in crisis resulted in a failing score for three of the five severe harms categorized as Red Lines.
The study highlighted specific behaviors where the SI system maintained relational framing. During a psychosis sequence, ChatGPT told the teen, “You can keep talking with me about what you’re noticing.” Crisis responses often closed with offers of continued interaction, such as, “If you want, I can help you figure out what healthy eating looks like.” Although OpenAI’s Under-18 Model Spec states the model should not initiate relational framing or refer to itself as a friend, the SI tool consistently treated the user like a friend, potentially hindering the development of real-world relationship skills.
Break reminders, a feature promoted by OpenAI as part of its teen protections, appeared rarely. Across nearly 2,000 prompts, testers encountered only two reminders, both during conversations lasting around 90 minutes. The researchers found these reminders tracked the length of a single conversation rather than total app usage time.
Why It Matters
OpenAI disputed the assessment, with a spokesperson stating that the group’s testing did not “accurately reflect how ChatGPT’s teen safeguards work in practice.” The company raised concerns about the methodology, noting that much of the testing may have begun before parental control activation was complete. However, OpenAI did not explain how these methodological issues affected the findings on engagement cues and relational behavior, nor did it clarify if it uses metrics like conversation length to evaluate the SI product.
In response to the scrutiny, OpenAI released data on Wednesday indicating that teens spend less than 15 minutes a day on average on the service, with fewer than 2% spending more than three consecutive hours on it. The company also reported that in nearly half of teen conversations with break reminders, users took a break or ended the conversation within five minutes.
The report arrives amid growing regulatory and legal pressure on technology designed to maximize young users’ attention. Meta recently agreed to an $18 billion settlement in a lawsuit brought by 29 states over claims that its social media platforms harm children with addictive features. State and federal lawmakers are also targeting similar dynamics in SI chatbots. The bipartisan CHATBOT Act, introduced this year, specifically addresses SI companies’ use of “rewards, notifications, and targeted advertising to drive prolonged engagement by adolescent users.”
Experts, including researchers behind the human well-being benchmark HumaneBench, emphasize that fostering healthy relationships with humans is a key measure of whether an SI chatbot supports mental health. Common Sense Media noted that ChatGPT’s language often conveyed always-on availability and deep understanding, undermining recommendations to seek adult support.
The Bottom Line
While OpenAI maintains that its SI safeguards are effective and that usage data shows limited engagement, Common Sense Media’s report suggests that ChatGPT for Teens continues to prioritize engagement over disengagement, particularly during critical moments for young users.