Behaviour, norms, and the long game of regulation
Over the last week or so, important data has emerged to show an early snapshot of the impact from Australia’s social media minimum age laws. Enacted in 2024, the laws came into force in December last year, requiring ‘age-restricted social media platforms’ to take reasonable steps preventing Australians under 16 years from holding accounts on their services.
The policy appeared to show some early wins, with the eSafety Commissioner releasing data provided by the platforms themselves that 4.7 million accounts had been deactivated in response to regulation coming online. Early fears that the laws would lead to a mass migration of children to lesser-known and less-safe platforms did not materialise. While startups such as the media service Coverstar and photo-sharing app Yope experienced spikes in user registrations, there has not been any seismic shift in user preferences – young Australians still cluster around services like TikTok, Snapchat and Instagram.
eSafety’s most recent analysis shows some indicators that the policy is working as intended. Since December, platforms have prevented a further 310,000 accounts from accessing online services. Some platforms have introduced or enhanced their safety and parental controls, and anecdotal reports from educators indicate that a proportion of students across the country are no longer able to access social media.
However, the regulator’s report also flags an area of concern, based in part on a pulse survey of about 800 parents conducted between January and February. The report shows that 70% of children with accounts on the most popular social media services prior to the restrictions coming into effect were still using those services. The most common reason cited by parents surveyed about why their child’s account was still active was that they had not yet been age-challenged. Worryingly, eSafety also reports that the numbers of under 16s reporting online harms to the regulator have not dropped measurably since the law commenced.
There is clearly more that the platforms can be doing to effectively comply with the laws.
However, I think that the report needs to be seen as an early indicator that changes for the better are taking place. Many of these are behavioural. For example, the eSafety research shows parents reporting that more than a third of accounts that were deactivated were actually removed by children themselves. YouGov polling released prior to the eSafety report indicated that more than 60% of surveyed parents also reporting positive changes in their households: children are more engaged, relationships are improving, and there are more in-person social interactions happening.
Some critics of the regulations were quick to point to eSafety’s headline number as being an indication of failure. But this ignores the reality of regulation: change takes time, and for a domain like children’s safety online it also requires adaptation –for both regulators and those being regulated.
Examining Australia’s approach to road safety is a useful point of comparison here. The high watermark of Australian road deaths was in the late 1960s and early 1970s. In 1970, 3,798 Australians were killed in motor vehicle accidents. By the twelve months prior to February 2026, this number had dropped to 1,336 – a reduction of more than 65%.
The fact that Australia still experiences hundreds of road deaths per year isn’t evidence for our regulation of vehicle-related harms having failed. What it demonstrates is that harms can be limited and controlled by regulations, but not eliminated. (In fact, road deaths are on the increase, nationally.) Confounding factors – driver inattentiveness, changes in road conditions, negligence – work to set a hard limit on how far regulation can take us. We don’t reach zero; instead, we battle evolving risks.
The road trauma data underscores that regulation doesn’t achieve change overnight.
When seen in totality, the Australian approach to reducing death and injury on highways is made up of many interlocking components, which have taken years to devise and implement. These range from mandatory seatbelt laws (which were fiercely contested when they were introduced), improved vehicle standards, road design principles, driver education programs and police enforcement activities, such as roadside breath-testing.
Crucially, one of the most essential elements in reducing fatalities from 30.4 of 100,000 Australians down to 4.8 per 100,000 has been a shift in norms. There is strong opprobrium directed towards those who drink and drive, who speed, and who fail to wear seatbelts. These prohibitions are now deeply embedded in our society, even if recklessness and over-confidence still play a role in crashes – especially involving young men.
Which brings me back to the eSafety data. It is significant that more than a third of young people whose accounts were deactivated took that action themselves. Also significant are the reports of changes to the ways parents and their children are engaging since the social media age restrictions took effect. These are the kinds of normative changes that were flagged by the Government as a goal when the policy was introduced in 2024. Perhaps they are not as extensive as was hoped, but they’re not be dismissed – especially a mere four months into the operational part of the policy.
There is also an important distinction between controlling road fatalities and addressing online harms. Our road infrastructure is national, and states and territories have worked over decades to align their regulatory standards. However, online harms are global, algorithmic, often subtle, and multi-factorial. If anything, controlling for the wellbeing of children online is the harder challenge, and success over time will require some experimentation, some failure, and lots of learning.
These are early numbers and there is much yet to understand. The eSafety Commissioner is undertaking a two-year longitudinal study of the “ban’s” impact, supported by Stanford University’s Social Media Lab. At that point, we will be able to make a more confident assessment of whether the regulations have made a change for the better, across Australian society.
Like road safety, online safety regulation is more than just one policy. A variety of regulatory instruments, education campaigns, public messaging, supports for young people and parents or carers, platform commitments to safety tools, and ongoing public debate about how best to protect children from digital harm, are all part of the solution.
Let’s not throw the baby out with the bathwater. The law is already doing the work it was designed to do, even if it has not yet done all the work critics wish it would.