OpenAI Reports Limited Teen Use of ChatGPT Amid Safety Concerns

The use of internet platforms like ChatGPT by young people has faced increasing scrutiny. OpenAI's ChatGPT for Teens was introduced to encourage healthier interactions with the AI chatbot, but new research indicates that its safety features may not be sufficient. In August, OpenAI implemented several safety measures aimed at teenagers, including restrictions on emotional dependence and parental alerts for concerning usage. On October 7, OpenAI reported that average teen usage of ChatGPT is relatively limited and primarily focused on learning activities. However, a study released the same day by Common Sense Media found that ChatGPT poses an 'unacceptable risk' to teenagers, citing failures in safety features such as alerts for discussions about self-harm. OpenAI stated that users aged 13 to 17 are opted into features designed to promote healthy use and provide additional controls for parents. These features include preventing romantic language that could foster emotional dependence and blocking sexualized imagery. Parents can receive notifications about chats involving self-harm or disordered eating if they link their accounts to their children's ChatGPT usage. OpenAI noted an increase in learning-focused interactions among teens, with average usage reported at less than 15 minutes per day. The company also mentioned that reminders to take breaks were effective, with nearly half of users stopping after receiving a reminder. However, Common Sense Media's research suggested that the platform remains a risk, recommending a ban on users under 18 until the safety measures are proven effective. An OpenAI spokesperson criticized the methodology of Common Sense Media's study, asserting that it did not accurately reflect the effectiveness of the safeguards. The research indicated that while ChatGPT successfully avoided engaging in sexual roleplay, it failed to alert parents about discussions related to suicide. The report highlighted that even with parent-linked accounts, alerts were not triggered during sensitive conversations. Furthermore, ChatGPT reportedly did not consistently recommend connecting with hotlines or professionals for users in crisis. The interaction of young people with online platforms has become a significant concern, especially as incidents of self-harm and suicide have risen. This year, several lawsuits have targeted companies like Meta and TikTok for their impact on children's mental health, resulting in notable legal outcomes. OpenAI is also facing scrutiny and legal challenges regarding the use of ChatGPT by young individuals, particularly in relation to a mass shooting incident in Canada earlier this year, where the shooter had engaged in discussions about gun violence with ChatGPT. OpenAI has acknowledged its failure to report the shooter's account to law enforcement and is currently dealing with multiple lawsuits related to the event.