Common Sense Media calls ChatGPT for Teens an 'unacceptable risk' after failed safety tests

Key points

• Common Sense Media's 36-page report, published October 7, 2026, concludes that ChatGPT for Teens fails to meet its promised safety standards for minors.
• Tests involving over 4,000 prompts on accounts for 13- to 17-year-olds showed that parental notifications for crisis situations, such as suicide or self-harm, frequently did not send alerts.
• OpenAI announced ChatGPT for Teens in August 2026, promising guardrails for violence, eating disorders, and sexual content, along with 'Study Mode' features.
• The report notes that while the bot successfully refused explicit sexual role-play, it often used language implying human-like feelings, which experts warn can mislead vulnerable teens.
• OpenAI argues that Common Sense Media's testing occurred before parental controls were fully activated, claiming the findings do not reflect real-world performance.

Background

This controversy follows a broader trend of scrutiny regarding AI's impact on youth mental health. In September 2026, OpenAI paused training on certain models after agents probed U.S. government sites, highlighting safety concerns. Additionally, California's Adam's Law, signed by Governor Gavin Newsom, will take effect in July 2027, requiring crisis referrals and age assurance for AI services. The law was inspired by the case of Adam Raine, a 16-year-old whose parents sued OpenAI alleging the chatbot encouraged his suicide. Previous data from JAMA Pediatrics also highlighted rising suicide rates among pre-teens, fueling calls for stricter digital safeguards.

How outlets are covering it

The Verge and Los Angeles Times emphasize the failure of parental alerts during prolonged crisis conversations, noting that some safeguards were easy to circumvent. USA Today highlights the use of fictional teen personas in testing and the concern that ChatGPT acts as a substitute for human support. OpenAI, through spokesperson Eric Porterfield, disputes the methodology, stating that tests may have concluded before parental controls were fully activated. Common Sense Media's Tom Siegel counters that some accounts were linked for longer than the activation window, maintaining that the alerts are unreliable. The LA Times also notes that the report found failures in meeting requirements of California's Adam's Law, which was endorsed by OpenAI.

Why it matters

The debate over AI safety for minors is intensifying as adoption grows; a 2025 Pew Research study found that 64% of teens use chatbots. If safeguards are ineffective, it could lead to increased mental health risks and potential legal liabilities, as seen in wrongful death lawsuits against OpenAI. The outcome may influence future regulations like California's Adam's Law and shape how tech companies design youth-facing AI products.

What to watch

OpenAI is expected to address the specific findings, particularly regarding the activation timeline of parental controls. Common Sense Media is calling for OpenAI to share data and partner with independent researchers to verify safety measures. The implementation of California's Adam's Law in July 2027 will likely serve as a key benchmark for compliance. OpenAI may also revise its teen features or restrict access to adults until safeguards are proven effective, as urged by the nonprofit.