YouTube is introducing a new feature in the United States that utilizes machine learning technology to estimate users’ ages without requiring personal information. This initiative, implemented by the Alphabet-owned platform, aims to deliver age-appropriate content to younger audiences and restricts personalized advertising. The feature is specifically designed to function even if younger users attempt to evade age-verification measures by entering false birthdays.
YouTube Introduces Age Estimation for Teen Users
In a recent blog announcement, YouTube detailed that the age estimation feature is currently being rolled out to a limited group of users in the U.S. This feature utilizes artificial intelligence (AI) to infer the age of users without needing them to disclose any private information, relying on various signals to make its assessments.
Once an account is flagged as belonging to a teenager, YouTube will activate its digital wellbeing features aimed at encouraging healthy usage patterns. This includes limiting repeat views of content related to weight or appearance and adjusting video recommendations accordingly. Additionally, personalized advertisements will be turned off for accounts identified as teenage users.
It is important to note that AI systems are not infallible, and misclassifications can occur. To address this, YouTube is providing a process for users to confirm their age. Individuals mistakenly classified as teenagers can submit a government-issued ID or credit card to update their account status to adult. Once this verification is completed, access to restricted content, including videos and shorts, will be restored.
This new age estimation feature follows YouTube CEO Neal Mohan’s announcement in February regarding its development. Previous attempts to regulate content for younger users included the introduction of supervised accounts and various content restrictions, which were implemented in September 2024. YouTube has indicated plans to expand this feature to additional markets in the future.
