Age verification: how Meta wants to recognize lying teenagers

Meta, the parent company of the social media platforms Facebook and Instagram, wants to use AI to determine the age of users in the future. Proprietary, AI-supported software is to divide users of the platform into two age groups in the future: older than 18 and younger than 18. According to the company's senior product manager for Youth and Social Impact, Allison Hartnett, this so-called Age Categorizer will be based on profile information and user behaviour to categorize user accounts accordingly. The software then scours the profile, the follower list and checks what content users interact with. According to the online medium Bloomberg , this also includes information such as birthday greetings from friends.

Automatic conversion to a teen account

If the software comes to the conclusion that a user must be younger than 18, their account is automatically converted to a teen account, regardless of the date of birth specified during registration. It is not yet known how reliable the Age Categorizer software is. However, according to a company spokesperson, anyone who has been wrongly classified as a teenager should be able to reverse this.

In September, Meta announced these so-called teen accounts for users under the age of 18. Owners of such accounts have limited contact options, for example. They can only be contacted by people they follow or with whom they have interacted. There are also restrictions on what content they can see. After 60 minutes of Instagram use, they receive a prompt to close the app and between 10 p.m. and 7 a.m. the app switches to sleep mode for them: notifications are muted during this time and direct messages are replied to via auto-reply. From the age of 16, some of the restrictions can be relaxed independently; users under the age of 16 require the consent of a parent or guardian.

Instagram and other platforms have long been criticized for doing too little to protect minors. In many places on the web, you can simply lie about your age to view adult content. However, according to Hartnett, teenagers who lie about this are less of a problem than expected.

Meta's AI-supported age verification is only possible because users reveal so much about themselves on the platform. According to the company, the most privacy-friendly solution to the problem of protecting minors would be for the app stores to take over age verification. However, Apple refuses – for data protection reasons. According to Bloomberg, a Google spokesperson said that there is no patent solution to the problem.

(kst)