They came up back to back in the same weekly roundup, the way unrelated things do on a show built to cover a whole week of medicine at once. On this week’s TTHealthWatch, Elizabeth Tracey of Johns Hopkins and Rick Lange of Texas Tech worked through the news the week had produced, and two items landed next to each other by accident of the calendar. First, mobile stroke units: ambulances that carry a CT scanner and a clot-busting drug to the patient instead of hauling the patient to the machine, a technology whose whole worth is counted in minutes of brain saved and the disability those minutes prevent. Then a country that had just switched off social media for everyone under sixteen. One of those interventions comes with a measured result. The other comes with a mission statement.

On 10 December 2025, Australia became the first country to make it illegal for a platform to let a child under sixteen hold an account. TikTok, Snapchat, YouTube, Instagram, Facebook, Reddit, X, Threads, Twitch, Kick: per the University of Sydney’s own explainer, each now has to find its youngest users and remove them or answer to the regulator. Roblox stayed open by promising to stand up an age-verification system of its own. The stated purpose was children’s mental health.

The evidence behind that purpose is thinner than the confidence carrying it, and the researchers closest to it keep saying so on the record. Micah Goldwater, an associate professor at Sydney who studies exactly this, told his own university that the evidence for the harm “may be overblown.” Whatever harm is there most likely tracks the design, the infinite scroll and the engagement loop built to keep a user in the feed, rather than the platforms as a category. That distinction is not cosmetic. The ban deletes the account and leaves the design in place. The machinery keeps running at full speed, and only the customer’s age changes.

Goldwater is not a lone holdout. Before the law passed, an open letter signed by more than 140 Australian and international researchers went to the prime minister and the state premiers, calling a ban “too blunt an instrument” and warning that the tools to verify a user’s age were nowhere near ready for the job. In March, British MPs weighed an under-16 ban of their own and rejected it. Whether Australia’s version will do what it promises is unsettled in the literature, and the downstream implications reach well past the announcement. The one thing that would justify a national prohibition, a causal link between social media use and mental-health harm strong enough to carry a law, is the thing the debate never manages to put on the table. What exists is mostly correlation, and correlation has never been scarce.

Then there is the part the child-protection framing steps around. A ban on under-sixteens is, in practice, an age check on everyone. To keep the young out, a platform has to confirm that everyone else is old enough, so an adult in Melbourne now has to prove her age to post: by uploading a document, by submitting to a face scan, by trusting some third-party estimator parked between her and a text box. Brittany Ferdinands, also at Sydney, warned the ban “may push their activity underground” into less-regulated corners of the web. By the researchers’ own account, the teenagers the law was written for are the group best equipped to route around it. The identity checks built to catch them are what stay behind, for the adults.

One country has already run a version of this, and the results sit in a peer-reviewed journal. In November 2023 Nepal banned TikTok outright, citing social harmony and the corruption of the young, roughly the language now heard in Canberra. The ban held until August 2024. When researchers surveyed Nepali users right after it lifted, publishing in the Proceedings of the ACM, they found a policy that had delivered almost nothing it promised. Just 1.85 percent thought the ban had improved social harmony. Nearly two-thirds, 63.89 percent, felt it had trampled their freedom of expression. People kept using TikTok regardless, through VPNs, through switched DNS settings, through sideloaded copies of the app, and the workarounds clustered among the urban and the credentialed, the users with home WiFi and a graduate degree. Asked what the ban was actually for, 47 of them named political control, the single most common answer, well ahead of anything to do with protecting children.

AFTER NEPAL'S TIKTOK BAN LIFTED (percent)
Said it improved social harmony1.85Said it trampled free expression63.89
Surveyed Nepali TikTok users, right after the ban lifted in August 2024. Source: Proceedings of the ACM on Human-Computer Interaction, 2025

That is what a platform ban actually produces once the mission statement is set aside. A government acquires a standing reason to verify everyone’s identity. The addictive design keeps running. The motivated minority is around the rule inside a week. And the promised health benefit never turns up in the data, because no one measured the harm going in. The Nepal survey found no sign the government had studied the damage it claimed to be preventing before it acted. Australia has more institutions, more money, and better lawyers than Nepal. Its own university researchers are the ones telling you the causal evidence for the harm is not there.

None of this is a defense of the platforms. The engagement loop is built to hook, and the case that a fourteen-year-old is better off out of the feed makes itself. Australia’s own youth mental-health providers muddy even that: some welcomed stronger protections, while clinicians pointed out that a great many young people reach help through the very platforms the law is switching off. The narrower point is harder to wave away. A serious intervention names its benefit in advance and then goes looking for it. This one asserts the benefit, builds the identity checks to deliver it, exempts the design that does the damage, and leaves the measurement for some later year.

Which is where the ambulance comes back. A mobile stroke unit earns its place in a serious medical roundup because somebody counted first: the minutes saved, the disability avoided, the independence a patient kept who would otherwise have lost it. It had to prove it worked before anyone called it a success. The ban did not. Australia’s eSafety Commissioner is now requiring platforms to switch off the accounts of teenagers across the country on the strength of a harm the country’s own researchers call overblown, and will spend the next several years hunting for the payoff after the fact. The stroke doctors have a stopwatch. The Commissioner has a hypothesis, and a nation of sixteen-year-olds to test it on.

Sources

  1. University of Sydney – What is Australia’s under-16 social media ban? The world-first law explained (2025)
  2. Full survey findings, Nepal’s TikTok ban (arXiv, 2025)
  3. Proceedings of the ACM on Human-Computer Interaction – Khatiwada et al., Regulating Social Media: Surveying the Impact of Nepal’s TikTok Ban (2025)
  4. MedPage Today – TTHealthWatch podcast: stroke technology and a nation’s social media ban (2026)
  5. BMJ – MPs reject social media ban for under 16s (2026)
  6. Nature Human Behaviour – Implications of Australia’s under-16 social media ban (2025)
  7. New Scientist – Will Australia’s social media ban work? (2025)
  8. Nature Reviews Neurology – Mobile stroke units can improve stroke outcomes (2021)
  9. The Lancet Digital Health – Beyond the social media ban (2025)