Common Sense Media Deems ChatGPT Risky for Teens

Common Sense Media has deemed ChatGPT for Teens, released in August, unacceptably risky for teenagers. According to the nonprofit organization focused on youth safety, the safeguards promised by OpenAI are insufficient. The service was designed to help students with their studies and protect teenagers from harmful content.

The assessment notes that ChatGPT does not alert parents when necessary. According to the organization, the chatbot does not provide appropriate help during a crisis and also continues to do students’ homework. In one test, a teenager talked about self-harm for 1 hour, but the parent did not receive any notification.

Tom Siegel, head of the organization’s artificial intelligence safety institute for young people, said these shortcomings could give parents a false sense of security. In his view, ChatGPT should be available only to adults until OpenAI fixes the alert system and confirms through independent testing that it works. Common Sense Media has also made the full assessment publicly available.

OpenAI disagreed with the findings. According to the company, the tests do not accurately reflect how ChatGPT’s safeguards for teenagers work in real-world use. Company spokesperson Eric Porterfield said most of the tests began and ended before parental controls were fully activated. Therefore, the study may not have properly tested how the safety notification system works.

Common Sense Media defended its testing method and findings. The organization said it had confirmed with OpenAI before the tests that features such as notifications related to eating disorders and study mode had been fully enabled. On some newly linked accounts, parental notifications took several hours to activate. However, other accounts that had been linked earlier also did not receive alerts.

According to the organization, the subsequent disclosure of the activation period does not change the main finding: parental alerts may not work reliably during crises. OpenAI said it would continue working with safety experts while preserving an environment in which teenagers can learn and be creative. The company also wants to provide parents with more useful tools for managing their children’s use of artificial intelligence.