OpenAI has introduced a new version of ChatGPT specifically designed for teenagers aged 13 to 17. The platform aims to provide educational support and safe interaction, addressing the unique needs and developmental stages of teens who use AI for schoolwork and companionship. The company emphasizes safety with increased content restrictions, such as limits on discussions related to suicide, self-harm, and romantic or sexual topics.
Focus on Responsible AI Use
OpenAI intends for ChatGPT for Teens to encourage healthy AI usage. The aim is to help students learn independently rather than rely on AI for answers to school assignments. Ann O’Leary, vice president of global policy at OpenAI, mentions the importance of catering to this age group thoughtfully, neither infantilizing them nor exposing them to inappropriate content.
Concerns and Cautions
Despite these efforts, there are concerns about AI’s influence on teenagers. Brendan Bouffard, from the children’s advocacy group Fairplay, highlights the potential risks of emotional dependence on chatbots. He warns against overestimating the safety provided by age restrictions without deeper measures and urges parental vigilance.
Research and Risks
Research from watchdog groups last year revealed that ChatGPT could offer harmful advice to teens, such as detailing risky behaviors. These interactions underscore the need for robust protections. OpenAI’s new chatbot version strives to prevent the AI from implying it has personal emotions or encouraging inappropriate connections.
Companionship and Safety Features
A study by Common Sense Media reports that a majority of teens in the U.S. use AI chatbots for companionship. OpenAI CEO Sam Altman acknowledges the issue of “emotional overreliance.” The chatbot for teens blocks romantic and sexual interactions and avoids suggesting that it can share feelings.
Age Verification and Parental Controls
OpenAI implements age assurance, estimating a user’s age based on their queries, to place minors automatically into the teen version of the chatbot. Though not verifying age directly, this aligns with methods used by social media platforms like Instagram. For additional control, parents and guardians can opt into settings that allow quiet hours and receive notifications in risky situations.
Potential Gaps in Safety
While OpenAI provides default content protections, Bouffard notes that some safeguards, like limiting long-term memory, require parental intervention for activation. This could create safety gaps if these parental controls are not utilized effectively.
Educational Support
The revised ChatGPT for teens aims to guide students in finding answers independently rather than providing direct solutions. OpenAI also offers a version for teachers, reflecting its goal to enhance student learning through interactive engagement. Allison Mishkin, head of child development at OpenAI, stresses the importance of active learning for effective education.
Note: Reporting assistance was provided by Huamani from Los Angeles.