Australia's own regulator found more than 81 per cent of under-16s still on social media after the ban

Australia switched on the world's first national social media minimum age on 10 December 2025. On Friday 31 July 2026 the regulator that administers it published the first before and after measurement of the law's effect, and almost nothing in it moved very far.

eSafety compared survey responses from more than 4,000 children and families collected before commencement with responses collected three months afterwards, in work the ABC reported was prepared by a global advisory panel of university researchers. Before the ban, nearly 86 per cent of the children surveyed said they used at least one age restricted platform. Three months later the figure was still above 81 per cent. Use described as daily or more often went from about 60 per cent to about 58 per cent. Account ownership was the one measure that moved by a double digit margin, falling from 52 per cent to 42 per cent.

The regulator's stated explanation was not that children had outwitted the age checks. Around half of the children who kept their accounts said the platform had never checked their age at all. eSafety said most under 16s who held accounts before commencement were able to keep them or create new ones, and gave the platforms' failure to implement effective age assurance measures as the main reason. The services covered were Meta's Facebook and Instagram, Snapchat, TikTok, YouTube and Reddit.

The figure most often repeated about the ban, 4.7 million accounts removed, deactivated or restricted, comes from an earlier document and a later date than it is usually given. Tech Policy Press, writing on 1 April 2026, attributed it to eSafety's Social Media Minimum Age Compliance Update of March 2026 and dated the count to mid January 2026, not to the first days of the ban. The same piece reported that about 70 per cent of surveyed parents said their child still had a Facebook, Instagram, Snapchat or TikTok account, and that nearly half said their child still had a YouTube account.

On wellbeing the evaluation reads closer to a null result than to either side's prediction. More than 65 per cent of the children said the ban had had no impact, the ABC reported, even though some of them had expected negative mental health effects before it started. Al Jazeera reported that the study found little change in physical activity, arts participation, family time or community engagement, and that parental awareness of children's habits went down rather than up.

The government's response has been to argue about enforcement rather than about design. Assistant Minister Andrew Leigh told Al Jazeera the law was, in his words, an important game changer in the conversation among parents, and said millions of accounts had been shut down. Two days before the evaluation appeared, the ABC reported that Communications Minister Anika Wells was pursuing a bill to double the maximum penalty for non-compliance to 99 million dollars and to give eSafety power to compel documents from platforms and third parties. The Age Verification Providers Association warned that fines could still be seen as a small price to pay if paying them helped make the ban look like a failure. The Law Council of Australia said the information gathering powers could inadvertently capture children and parents, and asked for the bill to be amended so that it better reflects what the government intends. A Senate inquiry was due to report in August.

Meta published its own account on 12 August 2026. The company says it removed more than 750,000 Australian accounts belonging to people under 16 across Facebook and Instagram as at 30 June 2026, that more than 500,000 of those came off before the law commenced in December 2025, and that it reached 1.3 million people with an education campaign in June and July. Those are the company's own figures, reported by the company. Meta also argues that age assurance belongs at the operating system or app store level rather than app by app, a position that moves the obligation onto Apple and Google.

What the evaluation does not settle is whether harm fell. The published coverage of the report describes no findings on cyberbullying or image based abuse complaint volumes, and eSafety's own publication could not be retrieved for this article. Nor do three months separate a law that is not being enforced from a law that does not work. The regulator's own diagnosis is that the age checks were largely absent, which means nobody has yet observed what a properly checked platform does to the numbers. Tech Policy Press counted comparable minimum age measures under way in the United Kingdom, Indonesia, Austria, Malaysia, Brazil and India, and consolidation of several proposed bans being considered by the European Commission. Those governments are legislating alongside a statute whose first evaluation mostly measured the conduct of platforms, and found it wanting, rather than the behaviour of children.