The 15-year-olds were far from happy when Australia banned them from social media. No more Insta, no more Snapchat, no more TikTok. From midnight December 10, 2025, tech companies that failed to prevent under-16s from having accounts risked fines of up to A$54.6 million. In the weeks before the ban, the teens shared tips on how to get around it (making a double chin to fool age-verification tools was one of them).
Only one in four planned to stop using social media once the ban came in. Two 15-year-olds, backed by a digital rights group, sought a High Court injunction to block it.
It’s easy to see why teenagers were appalled, says Jeff Hancock, a psychologist and founding director of the Stanford Social Media Lab in California, which is working with the Australian government to evaluate the effects of the ban. “They’re like, ‘Wait, I wasn’t doing anything wrong. I’ve built this world. I connect with my dance team and my soccer team, and I’ve created a small business, and it’s how I stay in touch with my grandma. And that’s getting taken away!'”
In this TV show, Australian parents and teenagers react to the country’s under-16 social media ban.
CREDIT: TODAY AND 9 NEWS AUSTRALIA
Teenage disgust Down Under has done little to slow global momentum in the ban’s wake, though. From Norway to Indonesia, Brazil to the United Kingdom, governments are crafting legislation to limit their teens’ access to social media — emboldened by widespread parental support, alarming declines in adolescent mental health and a best-selling book, The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness. And it’s not just politicians who are moving to take on tech. So are the courts: Meta, which owns Instagram, has been ordered to pay almost $1 billion in penalties so far in 2026 after US juries found the company liable for harming young users. Many more lawsuits are in the works.
Few dispute that some children are harmed by social media. Reports are legion of bullying and sexual and verbal abuse, and exposure to pornography, misogynistic and radicalization content and pro-eating-disorder material. Features such as infinite scroll encourage incessant time on social media that may displace activities that are good for well-being, such as sleep and face-to-face time with friends. Yet after two decades of research, evidence that social media is toxic to the mental health of most adolescents remains inconclusive.
Some hope that the natural before-and-after experiment presented by the Australian ban could be an opportunity to test the effects of social media more rigorously.
But there is also a growing call to move beyond the blunt tool of bans. Instead, politicians, researchers, advocacy groups and even adolescents are saying it’s high time to understand and regulate the algorithms and the experience they are delivering to teenagers, both of which remain frustratingly opaque.
“We need to pivot and ask how we can design these spaces to be safe, effective, engaging and supportive,” says Candice Odgers, an expert in adolescent mental health at the University of California, Irvine.
A vulnerable time
It’s easy to imagine how a poorly designed social media environment could damage a teen’s mental health. Each day, teens may spend four to five hours on social media that is largely invisible to their parents. And this is during a developmental period when risk-taking, as well as sensitivity to reward, peer judgment and social exclusion, is at a lifetime high and emotional control is not yet mature. Suggesting that social media might hijack these vulnerabilities, triggering mental health problems in the process, is not far-fetched.
But muddying the waters is the mix of harms and benefits that make up the social media experience. For every bullying post there is another offering support to marginalized teens. For every post promoting self-harm, there’s another focused on mental well-being; for every piece of brain rot, something educational, joyful, even transcendent.
And indeed, research to date has not delivered clarity. Some studies looking for associations between adolescent behavior and the hours spent on social media found that heavier use was linked to negative outcomes including depression and anxiety; poorer attention, decision-making and sleep; and suicide and body dissatisfaction. But these studies could not say whether heavier social media use is a cause or a consequence of the behaviors.
The debate has not been settled by longitudinal studies that have followed tens of thousands of teenagers in the UK, the US, Australia and elsewhere, sometimes for several years, to track how social media use early in adolescence affected mental health later. Some find no effect; others find small positive or negative effects. In a 2024 report, the US National Academies of Sciences, Engineering and Medicine documented in fine detail the harm that social media can cause to individual teens, while also acknowledging its benefits, and concluding that there was insufficient evidence to say that social media was causing adolescent ill health at the population level.
When the old ways don’t work
As findings from longitudinal reports continue to roll in, some experts argue that social media research is, worryingly, in a methodological crisis with existing studies unable to determine true causal effects. This is because they fail to recognize social media as a complex ecosystem of networks and feedback loops. Consequences of this complexity include new, emergent variables that confound conventional study designs. Like age, gender and social background, these variables should be corrected for, but they are not, says statistician Tobias Dahl of the University of Oslo.
To illustrate his point, Dahl and his colleagues propose a concept that most parents are intuitively aware of: the cost of missing out, or COMO. In a world where most teens are on social media, teens who are not using it may feel intensely left out. Any impacts of COMO on mental health would not be captured in the existing studies, potentially masking social media’s true impact. There is no quick fix, says Dahl. Simply trying to measure COMO and compensate for it, one runs into problems with other omitted variables, such as parents’ attitudes toward social media.
But what about a world where all teens are off social media? Aware that over 1 million young people are effectively part of a policy experiment, the Australian government’s eSafety Commissioner, which is tasked with enforcing the ban, is funding studies to assess its benefits or harms. The feat, involving surveys, focus groups and interviews; tracking social media use on phones; and over 4,000 adolescents and their parents and carers, is overseen by 11 experts, including psychologists, led by Stanford University’s Social Media Lab. It will look at impacts on well-being, including mental health, exposure to online harm (there is concern that the ban may push teens into the shadier parts of the internet) and effects on family life.
Three months in, the surveys indicated that adolescent sleep and mental health had not changed under the ban. But the full impact, if any, will not be clear until the youngest children have traversed their teenage years.
As they warned they would, most Australian adolescents have found ways around the new law, with more than 85 percent reporting using the banned platforms three months after its introduction.
Blame for the media giants
Of greater concern to the Australian government was a potential lack of compliance by the actual targets of the new law — the social media platforms themselves, which were required to “take reasonable steps” to prevent underage use. Social media companies, such as Snap and Meta, had previously argued that bans are a bad idea because of the technical challenges of age verification and the risks of children being driven to sites with fewer protections. But the Australian prime minister, Anthony Albanese, deemed that the platforms were not doing enough, and announced in June that his government would boost the eSafety Commissioner’s powers and double the fines to A$99 million.
Some researchers think that policymakers and researchers should change tack altogether. Instead of trying, most likely vainly, to implement wholesale bans, many are arguing that it is time to tighten control of the social media algorithms: the rules that dictate how content will be fed to users.
These algorithms are designed to keep adolescents glued to their feeds as long as possible by delivering content that may be benign or banal but is, too often, harmful. Every week, over 80 percent of Australians ages 17 to 24 see content they find potentially distressing, ranging from sexual violence to encouragement of self-harm, says Ella Moeck, an expert on emotional regulation at Adelaide University in Australia. That exposure is often unwanted and happens even when teens try to curate their feeds to avoid it.
Tech companies consider their algorithms proprietary, so researchers are faced with a black box when trying to understand how they operate and the experience they deliver to young people, says psychologist Susanne Schweizer at the University of New South Wales in Sydney. “If you are 15,” she says, “and you’re seeking information about depression, do you get it served up for 50 seconds? An hour? Three days? What is actually happening?”
Regulators are starting to push social media companies to provide transparency. For instance, the European Union’s Digital Services Act has since August 2023 required online platforms with more than 45 million users — which includes Instagram, TikTok and Snapchat — to share how their algorithms work, in a general way. But these measures do little to illuminate how the algorithms shape users’ actual experiences, says Daniel Angus, a computational social scientist from Queensland University of Technology in Brisbane.
Better tracking of actual use
And so, to crack open the black box, researchers are turning to tools such as the Effortless Assessment of Risk States app, which was originally intended to predict and prevent suicide, self-harm and other mental health crises. Once installed on a smartphone, EARS can track app use, including the tone of the user’s text messages, allowing researchers to explore social media use far more deeply. It’s still early days for EARS, but so far studies have got a more accurate handle on how long adolescents spend on their phones — five hours a day, which was two hours more than the adolescents reported.
This video from Australia’s 36 Months campaign described how the movement worked to achieve the first nationwide social media ban for under-16s.
CREDIT: 36 MONTHS
Another priority is uncovering what ads teens see in their ephemeral social media feeds; digital platforms make money from highly personalized advertising, sometimes delivered via influencers. To study this, Angus and his colleagues developed the Mobile Observation of Advertising Toolkit app, which automatically takes screenshots and analyzes them to pick out ads in people’s social media feeds. Through privacy filters, only the ads are sent to the researchers’ servers. “It means we can interrogate the ad ecosystem and see things we would not otherwise see,” says Angus.
A browser-based precursor to the Toolkit app identified illegal gambling ads and breaches of Australian political advertising rules in adult Facebook feeds, says Angus. The team is now using the Toolkit app to capture ads in the YouTube, Facebook, Instagram and TikTok feeds of young adults.
But could all this hoop-jumping to get a window on the dynamic, highly personalized interactions between algorithm and adolescent eventually be unnecessary? Tim Errington, chief research officer at the Center for Open Science, a US-based nonprofit organization, hopes so. The center has brokered an agreement in which tech giant Meta shares Instagram experience data with researchers, while itself remaining at arm’s length from study design and execution. The data-sharing process, which is managed by the Centre d’Accès Sécurisé aux Données in Paris, involves shipping a fingerprint-authenticated secure data vault to the researchers.
One study taking place under the agreement, which is known as the Instagram Data Access Pilot, examines the synchronous interplay between different types of Instagram activity and good and bad feelings in 18- to 25-year-old Australians, over 14 days. (The original plan was to enroll under-16s, but the Australian ban put the kibosh on that.)
“The important thing is that we are asking them right now, in the moment. We are not asking them to remember how they felt,” says Moeck, who is running the study. Participants complete two-minute surveys that pop up on their phones six times a day. Meta provides time-stamped activity data, including when posts, comments and likes occurred. Moeck hopes that the study will help to clarify how different types of behavior on social media affect emotions, and vice versa.
Other studies under the agreement will add more detail to the picture. One will investigate the impact of beneficial and harmful Instagram content on young adults’ well-being, another whether sensitivity to social media likes and other rewards is a risk factor for depression. The first published results are expected by June 2027.
Frustration on both sides
Research agreements like these are not easy to forge, says Errington. While researchers have been frustrated by not having access to data, social media companies are asking, “How do I know if I give you data, you’re going to do quality work? Or are you just going to sit there and find the storyline that you want?” he says. This way, he adds, “you hold both parties accountable to each other.”
Errington hopes that the Instagram Data Access Pilot will lead to an established framework for data-sharing between social media platforms and researchers. The sweet spot, he says, will be when the companies agree to test different algorithm settings in collaboration with researchers, so that the findings can directly inform policy.
For Angus, that day cannot come soon enough. “In the absence of transparency, we end up with really, really terrible policymaking,” he says. “There are many ways in which we could regulate constructively within the platform environment, rather than chucking kids off it.”
One option on the table is to independently regulate different systems within social media platforms instead of the platforms as a whole. This could mean requiring the recommender systems that choose which content to deliver to limit addictive features, or requiring search systems that expose young users to harmful content to be safe by design. It might outlaw features like infinite scroll. “We won’t only be preventing the harms we see now, but any other app that uses the same technology will be subject to that legislation,” says Sandersan Onie, a mental health researcher at Harvard Medical School in Boston, who has published a framework for such regulation.
Stay in the Know
Sign up for the Knowable Magazine newsletter today
New ways to regulate will likely be on the minds of policymakers around the world as they wait to see the impact of Australia’s social media experiment. Hancock urges patience. “This is not like you wave a wand, and all of a sudden, kids are out on the footy fields,” he says. Instead, an entirely fresh generation may need to navigate adolescence with limited exposure to social media before change happens.
But there is one way the Australian ban is having a more immediate impact. Across the globe, it has already shifted attitudes toward social media design and regulation, and elevated public scrutiny of the tools companies use to make their products irresistible. “It’s really changed the worldwide conversation,” says Hancock.
This article originally appeared in Knowable Magazine, an independent journalistic endeavor from Annual Reviews. Sign up for the newsletter.
