ChatGPT rolls out advanced age detection to protect ‘Under 18’ users from sensitive content

OpenAI is gradually enabling a new feature for ChatGPT designed to predict the age of its users, with additional safety features aimed at users who are likely under the age of 18. ChatGPT won’t prompt its users to share their age. Instead, ChatGPT will use AI-powered age prediction to determine if an account is likely held by a person under 18 and provide additional safety features automatically.

Privacy advocates have long pushed for companies to protect younger internet users, and this is one of the most important developments for AI safety online to date.

Why OpenAI Is Implementing Age Prediction?

OpenAI wants to keep teenagers safe from encountering unsafe content while using ChatGPT. There have been concerns over young people and generative AI, with worries centering around discussions around self-harm, violence, dangerous online challenges, and more.

ChatGPT’s Age Prediction Functionality

Instead of asking users to verify their age, ChatGPT will use predictive modeling to guess the age of its users. Some of the signals it uses include:

• Account age
• Peak usage times
• Patterns of usage
• Declared age at sign-up

If ChatGPT predicts someone is under 18 years old, the app will place users in “young viewer mode” by default. When in young viewer mode, ChatGPT will filter out content related to graphic violence, adult content, dangerous online challenges, restrictive dieting, and self-harm.

Misclassified Ages

Because age prediction is never 100%, some adults may find themselves in young viewer mode by mistake. If you are 18 or older, ChatGPT allows you to easily opt-out of the young viewer experience by verifying your age.

ChatGPT’s age verification process uses a third-party app called Persona. Persona verifies users’ ages by taking a selfie and government-issued ID if the user is in a country that requires ID for verification. After you submit your information, your additional safety features will be turned off.

Parental Controls

OpenAI also recommends parents and caregivers use parental controls if they want to customize how their teens use ChatGPT. Parents can set quiet hours and toggle settings like chat history (memory) on or off. Additional safety settings can also be customized by parents.

Future Developments

OpenAI will start rolling out age prediction features to users around the world over the next few weeks. The feature will launch in the European Union shortly due to local regulations. According to OpenAI, it will be fine-tuning age prediction “continuously, based on real-world experience and feedback from experts in youth safety.”