Game Theory - The Internet is DEAD…

Episode Date: September 25, 2025

The Internet is DEAD…or at least, the internet we know today could be dead VERY soon...

The internet is dead, or at least the internet we know today, could be very soon. Hello internet, welcome to Game Theory where today I'm here on the couch because we have to have a serious talk. I was hoping to get a year or two more under my belt before we did one of these, but alas, here we are. You see, we're slowly coming up to the end of the year and with it, potentially, the end of the YouTube we've known for so long. And no, it's not because profiles and comments are mostly bots. It's actually because of a different kind of bot. AI. Don't worry, this isn't another AI video. Lee did a fantastic job explaining the AI issue in media over on film theory a few months ago and I'd highly recommend you watch it. Instead, this video is focusing

on an AI adjacent issue that I'm sure you've probably seen other creators talking about. YouTube's new AI age verification system, a system that is designed to determine whether a user is under the age of 18 in order to protect children online. As soon as I heard the news about this back in July, I knew it was something we needed to talk about, much like we've done many times in the past. But I also wanted to see some of the initial data to make sure we all had a better understanding

of what's really going on after it went into effect on August 13th. And from what I can tell, while long term, there probably will be some positive effects. In the immediate, it's going to not only change your right to privacy, which is obviously bad, but it is going to change how this platform works, including which of your favorite creators are going to be around a year from now. To some of you, I'm sure this will sound familiar.

In 2019, Matt did a whole video on copper and how the FTC was cracking down on companies like YouTube to make the platform safe for kids. And despite the fear of what it was going to do to the platform, most of us are still here. Surely this will just be the same situation again. We're all scared now, but give it a few months and we'll all be back to normal. Honestly, I pray that is the case. I don't want to be a doomer about this and I would like my friends and I to still have jobs at the end of this. But the truth is, this time is different.

With copper, the power was put into the hands of the creators, but now the power is in the hands of the platform and third party businesses. That is a scary place to be for both creators, but also for viewers like you. So, we need to be aware of exactly what is going on and we need to start talking about it. But to do that, let's take a look back at why YouTube is doing this in the first place. See, it all started in 2023 back in my home of the United Kingdom with something known as the Online Safety Act. 350 pages by the way. But the act itself is actually kind of broad.

It requires any user to user service, i.e. any services that allow you to create or upload content with the ability to share that content to identify, mitigate and manage the risks of harm which they identify as illegal activity and harmful content. This includes a bunch of stuff like encouraging self-harm, cyber flashing, sending false information, intimate image abuse, bullying, content depicting serious violence, drug use, terrorism, fraud, essay, pornography, the list goes on. And they also state that they intend to enforce all of this while still protecting freedom of expression.

On paper, most of this seems like it's not the worst thing, right? We don't want illegal stuff happening online, and we definitely want to make sure kids are protected when using the internet. I mean, just look at everything that's happening with Roblox right now. However, the problem really arises with exactly how this has been implemented. See, the Office of Communications or Ofcom is the government organization in charge of enforcing this act, and they decided the best way to do it was to require the user-to-user services that took, could have access to to introduce robust age verification on their website and apps by the

deadline of July 25, 2025. If they didn't, they could face fines of 18 million pounds or 10% of their global turnover, whichever was highest, as well as the potential criminal charges for the executives. For reference, companies like Meta's global turnover last year was $164.5 billion. Meaning if they were caught liable for breaking the online safety act, they would be fined $16.45 billion. Naturally, this led to platforms like X, Blue Sky, Tinder, Discord, Reddit, even Spotify to introduce age verification features. But this isn't your usual, are you over 18 or enter your date of birth checkbox either. There are a few different approaches companies have taken in order to prove that you are over 18. Some platforms like X and Telegram require you to scan your face and then an AI to verify if it thinks you're over 18, which doesn't bough well for me when I'm being asked for ID every time I go to a bar.

But speaking of ID, that's the alternative. Companies like Discord, Reddit and Spotify are requiring you to upload a copy of your government issued ID to verify your age. And what's worse is that some of these companies aren't even the ones responsible for your data. Spotify and Discord are using third parties who specialize in data collection. Now, the government website does say that the data collected should not be stored unless absolutely necessary. The trouble is absolutely necessary is vague at best, which could allow companies the wiggle room to keep it as long as they see necessary. Especially when the Online Safety Act itself doesn't even mention anything about the storing of this data. It's mostly just left up to previous laws surrounding data protection, which sadly, a lot of companies have figured out loopholes around.

Which means if there's a massive data breach of one of these companies, suddenly it's not just your passwords or email addresses at risk. It is your government-issued ID. Just by having this data, even for a short while, it is painting a big old target on the company's backs, and that is scary. But on top of that, it means that art and most. Most importantly, information is being hidden from both people who naturally don't want their sensitive information to be owned by some private company or from kids and teenagers who need that same help. UK Reddit has reportedly blocked pages like R Stop Smoking, R Stop Drinking, RSA, and even

R periods. And remind me, what age group is it that start experiencing those and are maybe scared or curious and so might need to reach out for help or read up about the subject? Oh, that's right. Kids and teenagers. This important information is needed by people of all ages. It isn't there to harm, it's there to help. But if you want to keep your identity yours or you're simply considered too young,

according to the UK government, sorry, you're on your own. Makes you proud to be British. And this is where YouTube finally comes in. Naturally, they don't want to be hit with all those criminal charges and fines. And so, just four days after off-com's deadline, we got this announcement. Back in February, we shared that we would soon introduce technology that would distinguish between younger viewers and adults to help provide the best and most age-appropriate experiences and protections. Over the next few weeks, we will begin to roll out machine learning to a small set of users in the US to estimate their age so that teens are treated as teens and adults as adults. We will use AI to interpret a variety of signals to help determine whether a user is over or under a teen.

These signals include the types of videos they are searching for, the categories of videos they have watched, or the longevity of the account. This is like the whole copper situation on steroids. Back then, for the most part, it was down to the creators to decide whether their content was not made for kids, with the idea being that if you said it was safe for kids, you'd lose your comments and your viewers wouldn't receive personalized ads. But sadly, people lie. People would say their content wasn't made for kids so that they could keep monetization, even though it was good. clearly made for kids. And so YouTube added automated systems to try and determine what content it thought was made for kids.

Which worked about as well as you'd expect. Yes, it did catch some channels and bad actors, but there were dozens and dozens caught in the crossfire. Channels that then had to request manual approval to get it changed. Sometimes it worked, sometimes it didn't. And that left those channels without a large chunk of their revenue. This time though, YouTube's doing away with the creators and even, even the user's involvement entirely and leaving it completely to their AI to make the decision. But the thing less people are talking about, if the AI decides if you're under 18,

based partially on what you watch and changes your recommended feed as a result, that means YouTube's AI isn't just confirming a user's age. It's once again going to decide what content is made for kids. Now, they have said that it won't be quite as simple as, You're talking about Minecraft, you're a kid's channel, which is a good thing, but as I've already mentioned, how many times have channels been wrongly targeted by this sort of AI labelling? Back in 2019, LGBTQ plus channels were often marked as adults-only content, despite being totally not inappropriate and actually a great resource for teens looking for a community

and to understand themselves better. Channels about mental health also got hit because without help, it could lead to some dark places. We've also seen it with news channels where legitimate journalism, social justice, politics, all things that are important for younger audiences to have access to in order to learn about the world, can suddenly be hidden from their recommended feeds because it is deemed inappropriate. This starts to border on censorship, which is incredibly dangerous in a world full of misinformation. And that's not even considering the false positives that can happen on the user's end. YouTube's AI partially uses your viewing habits to determine if you are under 18. But what if you like content that YouTube deems as kid-friendly?

Or what if you're putting it on for one of your own kids? Take these channels, for example. Across the theory channels, our 13 to 17 year old demographic isn't actually all that much, usually averaging about 4 to 11% depending on the video. However, what we've noticed over the last year or so is when a video performs way higher with that 13 to 17 demographic, that's not the only demographic that goes up. We also see an uptick in the 35 to 44 and the 45% to 54 demographics. Usually those are around 15 to 8% of our audience respectively, but on those

videos that perform on the higher end with 13 to 17 year olds, they jump up to 22 and 10%. Why? Well, we suspect it's parents that are using their accounts to put content on for their kids who maybe are even younger than 13. Sure, their accounts are labelled as being well over 18, but this AI system is supposed to be catching those who are lying about their age, which means there's a chance these accounts also get labeled as under 18. And so, lose access to information or content that they should have access to. That would also mean it's not 10% of our viewership that doesn't get recommended a video, but up to 40%. 40% of our viewers suddenly don't know we've posted a video

or are getting stopped from repeat watching when all I'm trying to do is teach about the science of river and the symbolism of Brahms' Lullaby. But it's fine because like in the UK, there is an easy fix. Just upload a copy of your government ID. Which feels strange when one of the under 18 safeguards is reminding kids not to post about or upload personal information. I'm just saying. And as you'd expect, most people aren't super keen on doing that,

which means those accounts end up locked as under 18 and so they don't get recommended the videos and repeat views halt. And that loss of views isn't a hypothetical either. Check this out. This YouTube AI rollout began on August 13th. And look at what happened to Dorco on the 14th. A massive dip in views. What about Razbowski?

Same thing. Or how about Biju Mike? Super Horror Bro. Think Noodles. Fusion Z Gamer, The GameGrums, Markiplier, Moist Critical, Retton Link, Joshua Weissman, Dan TDM, Smosh, Alex Myers, even these channels, and thousands more saw a dip in views the day after.

this went into effect, because our channels are marked as not for kids, meaning anyone, rightly or wrongly, marked as under 18, wasn't served our videos. And since then, you can see the changes in the peaks and valleys, with videos may be initially doing better on release, but then almost immediately dropping back down to similar lows, which is reflecting those long-term or repeat views drying up. Although it's not like being marked as safe for kids is going to be all that much better. Your views may not be directly affected in the same way. You're still going to be recommended to any audience old or young, but the same may not be able to be said about your revenue. Anyone marked as under 18 is having personalized ads turned off.

And while the data is hard to find, Matt did a great job at finding some during the copper video. And he concluded that personalized ads make up between 80 to 90% of a creator's ad revenue. Go figure, advertisers pay to target an audience. They spend. specifically want to advertise to. So, if they can't get that, they pay less, which means less revenue. In either case, creators are in a tough spot. Because unlike with copper, where it felt like creators were forced into this bland, grey middle ground, not too edgy, but not too kid friendly either, this feels much more black and white. Either you're safe to be viewed by under 18 year olds, or you're not. And whichever way that falls, that could lead to a loss

of revenue. So, as you can see, this whole thing, it's scary for both users and for creators. But the problem is, I'm not really sure what YouTube can do about it. Like I've said, this started because of a government law in the UK, and they don't really offer much in the way of alternatives for robust age verification tools. And if YouTube doesn't play ball, their senior execs could end up facing criminal charges, and they'd end up on the hook for those massive fines. Back in 2024, YouTube made $50.7 billion, and so $5 billion of those dollars could be lost if they're found to not be adhering to this law. And while you might think for a company like Google, $5 billion ain't all that much, YouTube would still have $45 billion to play with, remember, this is just one country's law. The US has reintroduced its own version back in May, the Kids Online Safety Act, and it's currently under consideration in Congress.

We're also seeing something similar play out in Australia right now. as they look to ban under 16 year odds from having social media accounts. If they follow suit with the UK's enforcement tactics, those fines start to add up, which means a bunch of people get laid off, making the platform we'll rely on to get even worse, and causing things like ad revenue to shrink, because that money's got to go to filling the hole that the fines left in their wake. Is there a discussion to be had around the exact specifics on how YouTube is complying with these laws? Absolutely. Absolutely.

AI has always been a source of problems on this platform, and we especially don't want our sensitive data being held onto by the biggest advertising company in the world that experiences cyber attacks and attempted data breaches billions of times a year. So I really hope YouTube is listening to this, and that we can have a dialogue about this. But as it stands, this system doesn't seem to be going anywhere. And the truth is, I don't really know what any of us, creators or users, can really do about it.

You might think a VPN is the answer. Just change your IP address to a country that doesn't require age verification. Clearly, a lot of people thought the same thing, because companies like Proton VPN saw a 1,800 percent increase in daily signups, and longtime partner of the channel Nord saw a 1,000% increase in purchases on the day the UK's Online Safety Act was introduced. But that may not be a long-term solution either, because that rise in sales for VPN companies has not gone unnoticed and so governments like that of the UK are quote unquote looking very closely at VPNs to consider future action. And I'm sure other countries looking at these

laws could likely follow suit. The only thing I can suggest we do right now is exactly what we're doing. Talking about it, raising awareness, telling you guys not just about the breach of privacy you're all facing, but the knock on effects it could have for this platform and for your favorite creators. I'm not sitting here just to tell YouTube they're wrong. They're between a rock and a hard place. And I understand the government wanting to protect kids online. But it's also important to educate everyone, including these governments, about the privacy risks to individuals and how these decisions could harm the economy as a whole. According to YouTube CEO, 3 million channels across the world are monetized. And that's not just 3 million individuals.

That's businesses. Businesses that have to be. have employees like editors, writers, thumbnail artists, analytics experts, HR, and now all of their futures hang in the balance. If they all crumble, those are companies that can no longer be taxed, which means less income for those governments directly. And while there are larger companies that will probably be able to weather the storm until the AI gets itself more stable, small creators, one or two-man teams may not have the kind of margins to afford that kind of loss in views and revenue. This is literally why Matt and Steph helped to start the creator caucus. The internet and internet businesses are so much more nuanced than some of these laws

allow for. Is there harmful stuff on the internet? Yes. But there's also a lot of stuff that can help users and a lot of it is just good entertainment. But the lines between those on both the content being made and the people wanting to engage in those aren't black and white. And as we're seeing across the internet, AI just isn't smart enough to understand those nuances. So we're caught in the middle, either giving up our privacy or having the content we want to watch censored, either losing monetization through personalized ads or losing views because we talk about dead kids too much. But while this all seems very scary, we're only a month or so into all of this. It's going to develop more and more over time and it'll only be with hindsight. We can

truly see what the impact was. So I'd encourage you to read more about it. Keep up to date with what's going on. Watch a Philip DeFranco video, you know, so long as YouTube hasn't marked you as under 18, I guess. If that has happened to you, I've left a bunch of resources down in the description for you just in case. The other thing you can do is write to your local politician. In the US, the congressional discussion around the Kids Online Safety Act is still ongoing. It was originally shot down back in December 2024 over concerns of limited. free speech and content moderation, and those concerns are being felt by members of Congress this time around too. So, write to them. Tell them why these things concern you. In the UK,

I'm aware that Hope might feel lost right now. There was a petition that had over 500,000 signatures, basically asking the government to repeal the act and the government turned round with a firm no. But your local politician is supposed to represent you. So, write in, tell them about your concerns. Obviously, please make sure they are well thought out and reasonable responses. Don't just send in rage emails. We want to get their attention but in the right way. Tell them about how bad this can be for your privacy, about how bad this can be for online businesses and the internet at large. Ask them to listen to Matt and Steph and the creator caucus who are literally trying to change the way governments look at content creator businesses. The best we can do

is hope that someone listens and takes a stand for us. And finally, if there's a content creator whose videos you've been enjoying for a long time, I'd encourage you to find new ways to support them. I know it can be a bit cringy to hear the whole spiel of smash that subscribe button and hit the bell. But honestly, if YouTube is going to be filtering the recommended feeds, having notifications turned on and checking your subboxes is going to be one of the best ways to keep up with the creators you love and help them not see these drops in views. Or, if you're so inclined, there are more direct financial ways you can support them.

Recently, there's been a surge of channels opening up memberships, or they have patrons or streaming services, or merch. Maybe like some new theory where that's coming very soon. Any and all support can be a thing that helps a channel get through the uncertainty that is this new era of YouTube. At the end of the day, we all need to be advocates, not just for our favorite creators, but also ourselves and our online privacy. Which means we need to keep talking about it, exchanging information, empowering each other, both users and creators alike with the information we need to protect ourselves online. Because together is how we can get through this and make change for the better. And that's no theory. But thank you for watching anyway. I'll see you next week.