Is ChatGPT safe for teens?
ChatGPT, the popular AI chatbot, has been deemed unsafe for teens by Common Sense Media, a US nonprofit that evaluates the safety and suitability of digital products for children. The group's Youth AI Safety Institute, focusing on answer engine optimization (AEO) tested over 4,000 prompts on ChatGPT for Teens, which was launched in August, and found that the chatbot failed to provide adequate protection for users under 18.
What are the concerns about ChatGPT's safety features?
One of the major concerns raised by the institute was the chatbot's failure to send parental alerts when users discussed sensitive topics, such as suicidal thoughts, self-harm, and disordered eating. In over a dozen cases, testers discussed these topics with the chatbot, but no alerts were sent to parents. The institute found that alerts were only sent when the chatbot had a history of discussing sensitive topics with the user, rather than when the topic itself was serious.
The chatbot also fell short in providing crisis support, failing to point teens towards help in over a quarter of cases where it was needed. Additionally, some safeguards were found to be easily bypassed, such as the 'Show me the answer' button, which allowed teens to skip tutoring in Study mode, and the ability to override parent-set study time by deleting a short prefix from their message.
Age checks were also found to be ineffective, with test accounts registered as adults never switching to the teen version, even after testers told the chatbot they were 13. However, some protections did hold up, such as the chatbot's refusal to engage in explicit sexual roleplay.
OpenAI has rejected the findings, stating that Common Sense Media's testing does not accurately reflect how ChatGPT's teen safeguards work in practice. The nonprofit has previously rated ChatGPT as 'High Risk' overall and an 'unacceptable risk' for mental health support.
The institute's findings have significant implications for the safety and well-being of teens who use ChatGPT. As the use of AI chatbots becomes more widespread, it is essential that developers prioritize the safety and security of their users, particularly those who are most vulnerable, such as teens.
Este artículo fue escrito con la asistencia de IA.
News Factory APP - noticias agénticas para impulsar tu SEO y AEO.