Australia Under-16 Social Media Ban Was Defined by a Rule Written After It Started

Justice Jacqueline Gleeson wanted to know how much weight the examples could carry. Reddit’s barrister had offered the court teenagers who campaigned, reported and organised online. Gleeson asked what they proved about the ordinary teenager. Most young people, she noted, spend most of their time online doing other things. The exchange went to the centre of the Australia under-16 social media ban, which reached the High Court on Tuesday in a two-day hearing before all seven justices. Reddit’s written case describes the ban as denying young people “a key, but to a door that is already open.”

Parliament passed the ban in late November 2024, eight days after the bill was introduced. A Senate committee had 24 hours to take submissions. The law, Part 4A of the Online Safety Act, requires platforms to take reasonable steps to stop Australians under 16 holding accounts, and it puts no penalty on children or parents. Ten platforms were named: Facebook, Instagram, Threads, Snapchat, X, TikTok, YouTube, Kick, Reddit and Twitch, and the obligation began on 10 December 2025. The definition arrived later. In March 2026 the minister made Rule 4A. It added a further condition: a service is caught if it has a recommender feature, or a feature available only to logged-in users. The rule commenced on 26 March 2026, more than three months after the ban started. The eSafety Commissioner had told two parliamentary inquiries that a blunt-force approach risked shutting vulnerable young people off the support and affinity groups they relied on, and neither committee recommended a measure like the one Parliament passed.

Joanna Davidson SC, for the Commonwealth, told the court the ban answers seven credible risks posed by the design of the platforms, and that children’s developing brains were particularly susceptible to them. Parliament did not need further comprehensive proof to act, she said. The risks, as the Commonwealth’s written case lists them, run from addictive use and sleep disruption to depression, anxiety and disordered eating. The features said to drive them are the endless feed, autoplay, recommender systems, time-limited content, popularity counts and push notifications switched on by default. The Commonwealth’s answer to the alternatives contains a concession the challengers can use. Its written submissions accept that design-feature rules in their strongest enforceable form would require the same action as the ban itself. Platforms would have to stop children’s accounts or give children accounts with none of the features Rule 4A names.

eSafety, the regulator that enforces the ban, published its own before-and-after numbers in July, and they do not settle the argument for either side. Account ownership among children aged 10 to 15 fell from 52.4 per cent before the ban to 42.1 per cent three months after it began. Use of the restricted platforms in the previous four weeks fell only from 85.9 per cent to 81.5 per cent. Among the children who kept an account, the commonest reason was that the platform had simply not asked them to confirm their age. 50.2 per cent gave that answer, and 37.1 per cent had put their age down as 16 or older. Parents grew less aware of what their children were doing. The share who did not know their child had used social media rose from 23.3 per cent to 33.3 per cent. Use moved rather than stopped. Reddit, the platform bringing the case, rose from 6.1 per cent to 9.4 per cent, and Pinterest, Kick and BeReal rose too, while messaging and video games were used more often.

The filings argue longest about the age check. The Commonwealth’s own independent report, from August 2025, accepted that facial estimation was less accurate for non-Caucasian users. Indigenous populations were underrepresented in the testing, and skin tone was a known source of bias, with a mean absolute error of 2.76 years at age 13, 2.91 at 14 and 3.02 at 15. In March 2026, 438 scientists and researchers from 32 countries called for a moratorium on large-scale mandatory deployment. They said the technology was not ready and was relatively easy to bypass. The Commonwealth’s written answer is successive validation. More than one method would be used, checked again as a child nears 16, with liveness checks, VPN detection and monitoring for transferred accounts. Age could also be inferred from information already held on an account. A survey of 746 teenagers four months in, cited in the filings, found about 27 per cent of banned 14- and 15-year-olds complying. 63.8 per cent reported use of a banned platform in the past week, and three in four said getting around the checks was easy or very easy.

Reddit’s case says removing an account does not remove the content, and it removes the controls that came with the account. Most of what is posted on the restricted platforms can be read without logging in, and the features the Commonwealth blames, endless feeds and recommendations among them, keep working for a logged-out viewer. That covers Reddit, YouTube, Snapchat, Threads, Twitch, Kick, TikTok, Instagram and Facebook. An account holder could switch off recommendations and autoplay, set age-based limits on sensitive material, and use parental controls and screen-time limits. A child without an account loses those controls. The services where the filings say the worst bullying happens are outside the ban altogether. Messaging services and games, WhatsApp, Messenger, Discord, Roblox and Steam among them, are excluded. The agreed facts record that one child in three had their most recent or most serious cyberbullying experience on a messaging service. One in four had it on a gaming service.

Saree Boutros, commenting on a Facebook post by Senator Malcolm Roberts, wants the ban to go further. “You’re overlooking the positive impact this ban has had on some young people,” she wrote. “If anything, it needs strengthening, particularly with some parents still not buying in. I’d support extending restrictions to under-18s.” Vaughan O’Leary, on the Guardian’s Facebook post, took the political argument and turned it on the politicians: “Well technically that is true as politicians spend more time communicating via social media than any other source. So we have two options: 1) remove the ban 2) keep the ban and remove all politician social media accounts.” Daniel Geoffrey, on Senator Roberts’ Facebook post, went further in the same direction: “If you believe that young people deserve a voice, then allow them to vote from 16. Some of them pay taxes, are employed on full time basis, and at a minimum are affected by GST.” The filings support the premise behind both comments: teenagers use the services for news. The voting-age conclusion is Geoffrey’s own. A study of 1,027 Australians aged 10 to 17, cited in Reddit’s case, found 41 per cent had used social media for news the previous day. Of the 26 per cent significantly affected by the ban, 45 per cent said they had less opportunity to share views or take action. Gabriel Spacca, on the Guardian post, answered the enforcement question in a sentence: “The kids all know how to use a VPN. They’re ok.” Boutros wants the ban strengthened. The regulator’s evaluation recorded use barely moving and parents knowing less.

The alternative the challengers describe is already before Parliament. The digital duty of care bill, introduced to the House of Representatives on 15 October by Communications Minister Anika Wells, would require platforms to keep their systems safe proactively. Its September draft named five features to switch off for under-16s: algorithmic recommenders, logged-in-only functionality, endless feeds, engagement-feedback mechanics and disappearing content. Users aged 16 and over would choose between an algorithm-recommended feed and one showing only the accounts they follow. France’s under-15 ban was held unconstitutional in August 2026 by the Constitutional Council. It applied to services with no established risk, gave parents no way to lift or modify it, and assessed no individual child’s age, maturity or family circumstances. In the United States, the Supreme Court struck down a violent-video-games law in 2011 for want of proof that the games caused harm. The American groups NetChoice and the Electronic Frontier Foundation told the court that three of the four state feature laws cited in the filings were under injunction at the time.

No platform has been fined under the restrictions. eSafety has opened investigations into Facebook, Instagram, Snapchat, TikTok and YouTube over alleged non-compliance, and in October the minister said the regulator was preparing civil-penalty proceedings in the Federal Court.