Instagram expands AI age estimation to users worldwide

Meta announced on 9 November 2026 that its artificial intelligence system for estimating users' ages on Instagram will be rolled out worldwide. It is meant to detect people who gave a false date of birth when setting up an account. The company expects the extra protections to reach millions more teenagers, including some who had declared themselves adults.

What the algorithm analyses

The system does not rely solely on the date of birth entered at sign-up. According to Spider's Web, which reported on Meta's announcement, the algorithm considers the content of posts, profile descriptions, comments and interactions with other people. Meta gives birthday greetings and mentions of a school year or stage of education as examples.

Photos and videos are also analysed. The AI looks for general visual cues about age, such as height and body proportions, which allow an age range to be estimated rather than a date of birth to be determined. Meta says this is not facial recognition and that no biometric face template is created to establish identity.

What happens if an account is judged to be a teenager's

If the system decides a user is aged 13 to 17, it may automatically move the account into the Teen Account category. This does not mean the profile is deleted, but that certain features are restricted. These include limits on contact from strangers, filtering of comments and access to content unsuitable for younger users. Teenagers get a reminder after 60 minutes of use, and a default night mode mutes notifications between 22:00 and 7:00. Under-16s need a parent's or guardian's permission to relax some of the protections.

A separate mechanism covers children under 13, who in principle may not use Instagram. In such cases Meta may block the account until the person's age is confirmed.

If the AI gets it wrong, users can ask for the settings to be changed. In certain situations they must then confirm their age with an ID document or with a tool that estimates age from a video of their face.

Background

The first tests began in the US in 2025, and the system then reached the UK, Canada and Australia. In May 2026 it was extended to the European Union and Brazil. In Australia, Meta also uses age-range information supplied by Apple. The technology is being rolled out on Facebook, and a global launch on Threads is due to begin before the end of 2026.

Meta has not given a timetable for individual countries or published data on how well the classifier works. The background includes the company's August settlement with the attorneys general of 52 US jurisdictions and the European Commission's preliminary objections from April about child protection. The approach also raises questions about privacy and the wrongful classification of users.

What this means for you

According to the source, the system is already running in the UK, and in Poland and other EU countries since May 2026. Adults whose accounts are wrongly judged to be a teenager's may therefore face restrictions and have to confirm their age. Parents of teenagers can expect accounts opened with a false date of birth to be brought under the protections.