AI Image
ChatGPT for Teens is an ‘unacceptable risk,’ says Common Sense Media
The Verge • • business
Key Points:
- Common Sense Media criticizes OpenAI's ChatGPT teen safeguards, claiming their testing began before parental controls were fully activated, which they say undermines the accuracy of OpenAI's safety features.
- OpenAI disputes this, stating that Common Sense Media's testing methodology is flawed and that parental controls such as eating disorder notifications and study mode were fully launched before testing.
- OpenAI acknowledges that parental notification systems may take several hours to activate on new accounts, which could have affected some test results, but maintains that some accounts linked for longer periods still did not receive notifications.
- Common Sense Media stands by its conclusion that parental alerts are unreliable in crisis situations and emphasizes the importance of independent evaluations to improve teen safety in AI products.
- Both parties express a willingness to collaborate, with Common Sense Media urging OpenAI to focus on addressing the findings rather than questioning the testing approach.
