OpenAI has introduced a new system in ChatGPT that can predict users’ age based on behavioral data, without the need for follow-up questions. This will make it possible to reliably distinguish teenagers from adults, strengthening protection for minors in digital spaces.
If the system determines that an account belongs to a teenager, filters are automatically enabled — content involving violence, dangerous challenges, encouragement of self-harm, and unrealistic beauty standards is hidden. If the system makes an error, an adult user can complete a quick verification using a selfie, without uploading documents.
The restrictions were developed with the participation of experts in adolescent psychology and are based on current scientific research. The new feature is being rolled out in stages, including in the EU, where additional personal data protection requirements have been taken into account.
