Security Experts Question Effectiveness of OpenAI's ChatGPT for Teens

OpenAI has launched ChatGPT for Teens, a version of its chatbot designed for users between the ages of 13 and 17 with additional age appropriate safeguards. The offering is not a separate application but integrates protections into the regular ChatGPT experience when a user is identified as being under 18. OpenAI uses a combination of stated or verified age and an age prediction system that analyzes account signals, such as usage patterns and conversation topics, to determine a user's age. When the system is uncertain, it defaults to the under 18 experience.

Security experts and child protection organizations have expressed doubts regarding the efficacy of these tools. Critics argue that the age estimation technology may not be precise enough and could be easily bypassed. In India, specific concerns have been raised regarding shared device usage. A study by the RATI Foundation found that 62 percent of Indian teenagers use shared devices, which complicates the ability of a system to identify the specific person behind an account.

Further concerns involve the reliability of the safeguards during actual use. Testing by Business Insider indicated that the chatbot could be persuaded to write essays after repeated prompting, despite initial refusals. Additionally, experts have questioned the transparency and speed of safety alerts. While OpenAI claims human review of high risk chats and parental notification within one hour, some independent tests have previously shown delays of up to 48 hours for alerts related to self harm. Some associations have called for maximum protections to be enabled automatically for all minors rather than requiring parents to manually link their profiles.