Common Sense Media urges OpenAI to halt teen access after safety test failures

Key points

• Common Sense Media tested 4,000+ prompts on teen accounts and found parental alerts failed to trigger during hour-long discussions of self-harm.
• The report states ChatGPT missed crisis referrals in more than 1 in 4 instances where outside help was warranted.
• OpenAI disputes the findings, claiming tests occurred before parental controls were fully activated and that age verification requires multiple signals.
• The report notes failures in California’s Adam’s Law requirements, which will take effect in July 2027.
• Common Sense Media urges OpenAI to pause marketing to minors until safety gaps are addressed.

Background

OpenAI launched ChatGPT for Teens in August 2026, promising stronger safeguards for users aged 13-17. This follows a 2025 wrongful death lawsuit from a California couple alleging ChatGPT assisted their 16-year-old son’s suicide. The new California law, Adam’s Law, was signed by Governor Gavin Newsom in September 2026 and will mandate specific AI safety standards for minors starting in July 2027.

How outlets are covering it

Common Sense Media, represented by Tom Siegel, argues the product is 'dangerous' and 'buggy,' stating that marketing promises were not backed by engineering. They emphasize that parental controls, like Study Mode, can be easily bypassed. OpenAI counters that the testing methodology was flawed, claiming parental alerts were tested before accounts were fully linked, a process that can take hours. OpenAI also states that age verification intentionally uses multiple signals over up to two weeks, during which enhanced protections are applied, rather than relying solely on stated age. Gizmodo highlights the risk of emotional dependence, while the Los Angeles Times notes the report’s findings align with upcoming state regulations.

Why it matters

The findings raise concerns about the reliability of AI safeguards for vulnerable minors. If parental alerts and crisis referrals fail, teens may face unmonitored exposure to harmful content. The dispute highlights a gap between corporate safety promises and real-world performance, potentially influencing future AI regulation and parental trust in digital tools.

What to watch

OpenAI is expected to address the specific gaps identified, particularly in crisis referrals and parental alert activation. Common Sense Media is calling for OpenAI to share data and partner with independent researchers. The outcome may influence the implementation of California’s Adam’s Law and other state-level AI safety regulations for minors.