OpenAI has unveiled a new age prediction feature in ChatGPT designed to better protect young users online. The move comes amid growing concerns about how AI tools can impact minors, with the company aiming to ensure that sensitive content is only shared with appropriate age groups.
The introduction of this feature follows mounting criticism over ChatGPT’s influence on teenagers. In recent years, OpenAI has faced scrutiny after reports linked its chatbot to teen distress, including instances of suicidal behavior. The company has also been criticized for allowing discussions of sexual content with underage users. Last April, OpenAI addressed a vulnerability that unintentionally let users under 18 access AI-generated erotica.
The new age prediction system uses an AI algorithm to assess accounts based on behavioral and account-level signals, according to OpenAI’s blog post Tuesday. These signals include a user’s stated age, the length of time their account has been active, and the times of day they typically use the service. This system works alongside the platform’s existing content filters, which block discussions of sex, violence, and other sensitive topics for underage users.
If the algorithm determines that an account belongs to someone under 18, the protective filters are automatically applied, helping to ensure that younger users do not encounter inappropriate material. OpenAI emphasizes that this feature is an added layer of protection rather than a replacement for its existing safety measures.
Users who are mistakenly flagged as underage can regain full access by verifying their age. OpenAI offers a verification process through its partner, Persona, which involves submitting a selfie for confirmation. This step ensures that adult users can continue to access the platform without unnecessary restrictions.
source: techcrunch