Machine Learning to the Rescue

How Google Wants to Identify Minors

Google's app Family Link will get new features, including a new way to identify minors.

Published Modified

Google announces that during the year they will begin testing a new feature that uses machine learning to prevent children from accessing adult content on YouTube. This so-called "machine learning-based age estimation model" will be used to try to predict if a user is under 18 years old, in which case an appropriate age filter will be activated on the account. 

To make these assessments, YouTube will analyse data such as the user's searches, video categories, and the age of the user account. If the model indicates that someone is a minor, the platform will apply its standard age filters, which include blocking inappropriate videos and activating the SafeSearch filter. 

The test phase for this new filtering tool starts at the end of the year, and a global launch is planned for 2026. The feature will also be tested on other Google platforms, indicating a broader strategy to protect young users on the company's services.