In the spring of 2027, social media will be banned for under 16s in the UK. How will they enforce this? Ofcom have some ideas, and have just released a rapid assessment of age assurance methods, which I would like to discuss here. But first I have a question: have you ever read Freakonomics? The book that cherry-picks data to make wild claims about the economic and social phenomena of modern life? In it Levitt & Dubner extrapolate that the legalisation of abortions in the 1960s was the direct cause of crime reduction in New York in the 1990s. Someone at economics school forgot to tell them that thing about correlation and causation. Are they saying that all the people who would have become criminals in the 90s were aborted 20-30 years prior? That all unwanted children become drug dealers or petty thieves? The book is a nightmare. They also say that residential pools in the US are more dangerous to children than a gun, in a home that has both. Something must happen to your brain when you sit and look at statistics all day; you forget that a gun is a lethal weapon literally designed to kill people, and a pool is a recreational vat of water in which unfortunate incidents sometimes happen. Why bother even comparing the two? Do the authors really love guns THAT much?
Freakonomics is a pretty old book now, so hopefully it can’t hurt us with its nonsense any more. Thank god it was written before social media started its blind rampage against the wellbeing of children. Imagine the parallels they would have drawn nowadays, in a world where the common consensus is that TikTok turns children into gay communists. Or maybe that social media is as dangerous as cigarettes! Even though social media — an amorphous thing that happens on a screen, for free — is obviously nothing like a pack of expensive tobacco sticks that you buy from a shop and light on fire. Social media comes with unseen and immeasurable benefits; it’s hard to quantify the value of the warm feeling you get when you know that you have found your people in a niche online space. With cigarettes however, no matter how cool you look while doing it, inhaling smoke into your lungs is always bad for you in every possible way. Social media is also not ‘addictive’ in the same way cigarettes are. Everyone responds differently to those specially designed low-effort dopamine hits: e.g. Elon Musk got so addicted to Twitter during covid that he bought it. That would never happen to me, I don’t have an addictive personality and I also am not a billionaire supervillain. I also don’t like social media any more; I enjoy trying to understand why people use it in the ways that they do, but I miss the days when we were all arguing about whether or not you wash your legs in the shower. Rather than now, when we argue over whether something is AI generated — or just film ourselves stripping copper wire out of the walls.
Just like the authors of Freakonomics do with, I guess, the whole of modern society, those lobbying for a ban totally misunderstand the features of our online world. If we follow the logic that all social media is both addicting and dangerous, all online safety efforts have to be geared towards making it safe (or making it not exist), rather than making it good. But the big main social media platforms don’t represent the whole internet; there are social spaces out there, both good and bad, that don’t run on newsfeeds optimised for engagement. What about AO3? There are no algorithms there, just thousands and thousands of words of fan-fiction that you have to search for manually — which is fine because AO3 users know EXACTLY what they are looking for on arrival. Go and see for yourself! It cascades with unhinged meta-dramas and alternate universes based on canonical universes you’ve probably never even heard of. Except you can’t see it, not right now. The site fell victim to one of the five million cyberattacks that happened in October, thanks to those other people who also hate the web (AI companies). AO3 users are starting literal support groups over this. The site will be down for 2-4 weeks. The fanatics can’t wait that long. One user has said that they are “starting to get the shakes from Dramione withdrawals and whofic doesn’t hit the same anymore.” They are all rabidly addicted to writing and ingesting fan-fiction — an algorithm didn’t do this.
AO3 is addictive because it’s entertaining, and finding ways to be entertained is only getting harder, especially when living under the greedy succubus that is the current media industry. Do you know what a lolcow is? It’s when baiting and laughing at someone online is more fun than watching a Netflix original, and it’s a culture that transcends any one platform. You can ban kids from social media, but you can’t protect who the next lolcow is going to be. People have been doing this for years and seemingly can’t stop, suggesting it probably has some addictive qualities, and it has nothing to do with algorithms. Websites like Kiwi Farms or 8kun are privately run message boards which flow thick with slurs, targeted attacks, and self-hatred. The users might rally around ruining someone’s life for fun, or just ruining each others’ with intense, abject nihilism. The sites themselves are not designed to scale up and don’t rely on any kind of algorithmic gamification; they are fringe websites that intentionally operate outside of the mainstream gaze, because they do not want to be perceived, disturbed, or caught. Consider these the ‘underground’ of online social spaces. Now think about this: when you ban something, what happens? The activity doesn’t stop, it just goes where enforcers cannot see it. When a 15-year-old boy can’t watch Clavicular streams on Kick any more, maybe he’ll just spend more time hanging out on looksmaxxing forums, which I’m sure are nowhere near as deranged! If safety is the main concern here, a ban makes no sense at all.
In Design Justice, Sasha Constanza-Chock discusses designing for the fringes to avoid a monoculture of services that only centre one subset of lived experiences. Mainstream platforms like Meta, Twitter, etc don’t follow this principle at all. Why would they? What they’ve done has made a small group of white men very rich. But their homogeneous approach has lead to bizarre design choices such as a ‘real name’ policy, which forced trans people to use their deadnames, or flagged indigenous names as made up. Or the practice of outsourcing content moderators to one locale, who then cannot make effective moderation decisions for all contexts, cultures, and languages, resulting in dangerous outcomes. This is exactly why designing for edge-cases is so important: because the edges are where the most precarious, vulnerable, and marginalised tend to be. It’s also why places like AO3 or 8kun exist — whether good or bad, they’ve designed spaces that work for them, where they can express themselves without judgement. Kiwi Farms is a forum that was originally spun-off from Encyclopaedia Dramatica, in order to torture just ONE person. Even though it’s horrible, it’s also a truly niche purpose-built community-led project. To give a slightly more positive example of this: there is a library that exists in Minecraft which publishes news and information from certain countries whose governments would have either censored or manipulated the truth. These are the lengths that communities go to in order to maintain an internet that is as free and open as it possibly can be for them.
So let’s fast forward to when the government have decided exactly what they’re going to ban — and not build — for children. How does enforcement work? And if we are banning, rather than trying to make something better, should we not at least design those enforcement methods for the fringes? Shouldn’t they be thinking about how teens can get around age-gated apps, or how they may stray to weird little alternative spaces that fester in the dark, who’s owners have no interest safeguarding? Teenagers are already repurposing Spotify comments to chat to each other, and that’s just a small case, so what’s going to happen when we get a nationwide ban? Maybe the Yelp reviews for Bumps Centre International (it’s a dermatologist based in Deptford) will become the next place for kids to share affiliated links to discounted Clickeez — and maybe the Clickeez Amazon reviews will be where teens ask each other for relationship advice. They can’t age-gate everything!
And I really hope they don’t, because the options laid out by Ofcom sound like a privacy nightmare. One of them is photo ID matching. This requires uploading a passport or drivers licence, and matching it with a selfie. Platforms will decide whether they will use a third party, or go ahead and scan children’s faces and IDs themselves. Photo ID matching is considered one of the best options, because it works as a fallback to facial age estimation, where facial recognition technology tries to guess the age of the user. Who is collecting this data and how will it be stored? Maybe avoid whoever Discord was using; there was a breach that exposed 70k government IDs in 2025. A sign of things to come! Speaking of ID, there are over 100k teens in the 16-17 range who don’t have it, so will effectively be locked out if they don’t pass any of the other checks. Something else we need to consider is that this all counts for adults too. If you don’t want to get your face or ID scanned every time you check the Cyberpunk subreddit, you can either give up, or just find a website that uses one of the other methods, such as checking if your email activity looks grown-up enough, or seeing if you use the parental controls provided by your mobile network. Even without thinking about the very easy circumvention techniques available (e.g. Firefox is a free web browser and comes with its own VPN lol), these age assurance methods just don’t seem worth enforcing a ban that already looks like it won’t do anything. Kids will find a way around it, meanwhile adults have to go through an ID check every time they want to discuss their favourite video game or, heaven forbid, maybe watch some p*rn.
From my perspective, a nationwide ban of a group of services which are useful and fun feels very authoritarian. Time spent fiddling around with age assurance is time not spent making the internet better, and working with both children AND adults to build something that doesn’t reward us for becoming outraged or bait us into being laughed at by Twitch streamers. I’m not sure adding the friction of an ID check is really going to change any of this.
Maybe someone will change my mind at the event I’m running tomorrow about the ban. Hopefully see you there! If you weren’t aware of this event: I’m sorry I think we’re at capacity, but there’s always next time, so follow the calendar on Luma to be among the first to be invited in the future.