OpenAI has initiated the introduction of an age estimation feature for ChatGPT to detect accounts that are likely owned by individuals under the age of 18 and implement extra safety measures. The new system, as announced by OpenAI, will use various factors like account duration, typical online activity times, usage habits, and the user’s self-reported age to estimate the user’s age. In cases where an account is identified as potentially belonging to a minor, ChatGPT will automatically limit access to certain sensitive content categories, such as graphic violence, sexual role-play, depictions of self-harm, and content that promotes extreme beauty standards or dangerous viral challenges.
Users who are mistakenly categorized as under 18 can confirm their age by submitting a selfie through the Persona identity service to regain full access. Additionally, parents have the option to utilize extra parental controls, including setting specific quiet hours or disabling particular functionalities.
The deployment of this feature is currently in progress for consumer plans, with plans to implement it in the European Union in the upcoming weeks to adhere to regional regulations. OpenAI mentioned that the methodology is based on insights from child development professionals and that they will continue refining the model’s accuracy.