How will Australia enforce its new social-media law?
In November 2024, Australia passed a law that will ban children under the age of 16 from social media platforms. It’s the first law of its kind in the world and it had broad support in Australia. Over three-quarters of the public supports the law. The prime minister said he wants children to have a childhood again.
The law is scheduled to go into effect later in 2025. But it’s still not clear exactly how this law will work. In this story, we’ll talk about two big questions that social media platforms and the Australian government will have to deal with. First, how are the platforms supposed to verify the ages of their users? Second, what even counts as a social media platform?
Start with age verification. The law says that social media platforms must take “reasonable measures” to verify the ages of their users. But this is vague; there are several ways to verify age online, but each way has its problems.
The easiest way to verify a user’s age online is with a credit card: enter the credit card number and the site can check the user’s age against the bank’s records. But this is unreliable because it’s so easy to borrow or steal another person’s card.
Another way to verify ages is with government identification. Upload a copy of your identification and then take a picture of yourself with your device. The system attempts to match the photo on your ID with the photo you take in the moment. If your photo matches, and if you’re of age, then you’re approved. This method is more reliable, but it’s expensive. There many, many hundreds of types of government ID: platforms need to separate the real ones from the fake ones.
But here’s another thing. Some teenagers, even those past their sixteenth birthdays, might not even have an official government ID. I didn’t get mine until past my sixteenth birthday. So many sixteen-year-olds in Australia won’t be able to prove their ages using only government ID.
That leads us to the newest way to verify ages: artificial intelligence. Take a picture of yourself and AI will estimate your age. This method is new and improving, but not perfect. Systems can make a reasonable guess at users’ ages, accurate to within a couple of years.
But the accuracy is uneven. It works best for people of middle age and with lighter skin tones because those faces are over-represented in the models’ training data. But the system works less well for people outside middle age and for people of other ethnicities. So AI age estimation can’t be the only answer, either: the whole point is to verify users around age 16, and the models aren’t good at that yet.
The question is, how will the Australian government and social media platforms agree on what is a “reasonable” effort to verify ages? And how easy will it be for kids to get around the new verification systems?
The second question about Australia’s law is, “what is a social media platform?” This is more complicated than you might think. The law says that a social media platform is one that facilitates social interaction, content-sharing, and user-generated content.
The problem is, there isn’t a bright line that divides social media platforms from other types of sites. Many sites have social elements even if they are primarily for other things. Australia’s eSafety Commissioner will be in charge of determining which sites are classified as social media and which are not.
But let’s have some fun. Remember that the age-verification law applies to platforms that facilitate social interaction, content-sharing, and user-generated content. That’s the definition. So, Facebook, X, Instagram, and TikTok all count. They are primarily about user-generated content.
Messaging apps like WhatsApp and Telegram are a borderline case. They primarily offer direct messaging, calls, and file sharing between two people. They also have group messaging for small groups. That wouldn’t seem to meet the definition of “user-generated content.”
But these platforms also have stories features, which are definitely social. Telegram is known for its large groups and broadcast channels, which feel like a private, closed social network. It’s not clear if WhatsApp or Telegram will be subject to the new law.
Twitch is a live gaming platform. Content creators can stream themselves playing online games, or simply just living their lives. But the streams also have a live comment feed, which can feel social. Something tells me Twitch and similar platforms will have to comply with the new law.
Next, what about a site like Goodreads? Goodreads is a platform that lets you look up books and discover new books. But there are also a lot of user-generated reviews. You can create a profile with an activity stream, and connect with other users. Will Australia really kick 14-year-old book nerds off Goodreads? For that matter, if book reviews count as social, what about shopping sites like Amazon, which have lots of user-generated reviews?
By the same token, Strava is a fitness tracking app. But it also lets you connect with friends and post your workout details, including the route you took on a bike ride or a run. Is that what Australia means by harmful social media?
What this shows is that “social” is not a place that you go or an app that you open. It’s a characteristic of online life. People like sharing their opinions, generating content, and creating profiles. It’s built into so many of our online experiences.
Australia wants to keep kids off “social media” platforms. But so many sites that kids and adults use have a social component. I wouldn’t want to be that eSafety commissioner, trying to decide which platforms are subject to the law and which are not.
Jeff’s take
This is a law that might have good intentions behind it, but I really don’t think the lawmakers thought this through. This is going to be impossible to enforce. They worry about the cyberbullying and body shaming and pro-suicide messages that can show up in the TikTok or Instagram algorithms, and that is a legitimate thing to worry about. But is that really a problem on LinkedIn? And should every app with even a small social component now have to collect personal data on your age?
It will be interesting to see how this plays out between now and the time the law comes into effect.
Great stories make learning English fun