That of verifying the age of users, that is people who log in or register to the various services, it will be one of the fundamental games of the Internet of the future. Or even already in recent years, judging by them interventions by the authorities in this sense both in Europe and in the United States.
That the various platforms, especially the largest and most important, are somehow able to monitor the age of their subscribers is in fact the main means through which children can be protected onlineprotecting them from content and information that perhaps they are not yet ready for or even from malicious people who pretend to be their peers even if they are not their peers.
It is a question that also concerns artificial intelligence (ChatGPT was blocked in Italy for a month precisely for this reason and Google does not yet make Gemini accessible to minors) and also the sites that collect material for adults, which is the reason why in this test we have also including PornHub. What we did was use a computer that wasn’t ours and a browser that didn’t already know us to try to register on the main social networks, on Google, on YouTube and indeed on ChatGPT and PornHub, pretending to be an 11 year old who would lie about his age. Only afterwards did we ask the various platforms for confirmation of what happened and what their procedures were for so-called age gating. Let’s see how it went.
Social statistics and generation gap, the YouTube case: 18 year olds like it and also those over 65 by Emanuele Capone 07 February 2024
Facebook and Instagram
On the main Meta social networks, the first filter is the one that almost everyone uses: your age is asked during registration. It seems silly (but as we will see there are sites that don’t even do this), but it’s been around for a while the industry standard and it is what is recognized as lawful and acceptable even by the legislators of many countries. The app does further follow-up checks if it detects or becomes aware of what it calls “discordant signals”. For example: if the date of birth is changed after opening the profile, if a person has adult friends and adult interests but presents themselves as 16 years old (or vice versa) or if other users report the profile as under age, a request for documents and to undergo a features check is triggered through Yoti, an age verification service based on face scanning and AI and which works very well (And also used by some adult sites).
Google, Gmail e YouTube
A single procedure for everyone, given that the account is the same for using the various services. When you create a new profile you are asked if it will be for an adult or for a child: you can lie (which is what the imaginary 11 year old in our test did) or go through the procedure that makes the adult the supervisor of the minor , using the helpful Family Link app. Those who decide to pretend to be older have immediate access to everything, but “if we become aware that you may not be the minimum age required, you will have 14 days to update your account to meet the age requirements, otherwise your account will be deactivated”. To confirm your age you need an identity document or a credit card, the data of which is stored “only for the time necessary to satisfy legal requirements”.
TikTok
To open a profile you need to enter a date of birth: you can lie, but TikTok does some things that competitors don’t. Simply put, check what is published by the account and verify that there is consistency with the declared age: the analysis of the videos, comments, the use of facial recognition and any reports from other users are also collected. In our many years of experience on the app, it is very difficult to escape these checks: if anomalies are found and there is a suspicion that the new account belongs to a person under 13 years of age, the presentation of an identity document, using a credit card, or even sending a selfie with a parent or legal guardian. If there is confirmation that the user is under 13 years of age, the account is first suspended and then banned: in the last quarter of 2023 alone, the platform removed almost 20 million new subscribers worldwide for this reason .
At the bottom right, the age verification banner that appears on ChatGPT
ChatGPT
After the aforementioned scolding from the Guarantor in 2023, the best-known conversational AI has implemented a real age verification tool (always using Yoti) which should prevent children from accessing its services. And yet, there is a however: in our test we found that you can lie and that you have 60 days to undergo the check. I am a good two monthsin which our imaginary 11-year-old could have asked ChatGPT dozens and dozens of questions on the most disparate topics, even on topics that perhaps he would have been better off addressing with a parent or guardian, receiving the not always adequate answers typical of artificial intelligences.
On Elon Musk’s social network, which is now called there is then no subsequent control to understand if you are lying or not. The only thing the site bothers to verify is that the new user is not a bot, with a complicated, boring and long question and answer process to confirm that they are human. Once you pass it, you have access to any of the things you find on Twitter, from porn to fake news and pseudoscientific information of dubious foundation, such as those on flat-eartherism, on the dangers of vaccines, recommendations to watch an eclipse with the naked eye and so on. And no one will check whether you are really of the right age to read them.
PornHub
Every time you log in, the most popular adult site asks you to tick the box I am 18 years old or older which dates back to the prehistory of the Internet (image at top of page). It is objectively little, but at the moment it is the only barrier against illegitimate access.
Social network TikTok suspends the Lite rewards app after the EU accusations: “It can create addiction” by Emanuele Capone 23 April 2024
A problem (especially) for Apple and Google
Based on both the short tests described on this page and the last 30 years spent online, it is clear that the problem of age verification is far from solved. And also that there are some who deal with it more correctly than others. In a not perfect but more concrete way, that is.
Having to summarize, TikTok and Instagram are largely the platforms that do best (with a slight advantage for ByteDance) while Twitter and PornHub are the worst, not surprisingly for both. At the same time, Google it seemed to us to be the one that has the greatest potential to be effective: given how pervasive and everywhere Mountain View’s services are, and given the widespread and very likely use of artificial intelligence, it is difficult not to get caught if you lie about your age.
Find a solution will not be easy, but perhaps we are looking at the problem from the wrong point of view: we should not demand controls from apps or platforms, but from those who make them available. From the Apple and Google stores, that is. It is a discussion that (for example) Meta has been pursuing for some time, as it recently reiterated Antigone Davis, one of its vice presidents (Who), and it makes sense. As for us users, do we really want very private information such as that on credit cards or our documents, including numbers, photos and security codes, to end up in the hands of employees of Facebook, Twitter, TikTok and the like? This is sensitive data, which social media doesn’t even want because for them it translates into other problems: How should they treat them? How should they store them? How long should they keep them? For Apple and Google this is theoretically easier, as Davis recalled: “We are pushing for age verification at the operating system level, in digital stores. Teenagers can only buy a smartphone with parental consent, therefore manufacturers verify the age at the time of purchase and have this information as soon as the phone is set up. If they were shared with apps, we would no longer need to collect data.”
In simple words, and returning to the imaginary 11 year old of our test: if his parents bought him a smartphone and set it up like the smartphone of an 11 year old, this information would be stored in iOS or Android (depending on the type of phone) and therefore also within the respective App Store or Play Store. What if that eleven year old tried to download an app for which he is not yet of age, the download could easily be prevented. Solving the problem at the root. If Apple and Google wanted to take charge of this too, obviously.
@capoema