eSafety's own report shows five million accounts gone and most under 16s still on social media

Australia's social media minimum age obligation took effect on 10 December 2025. Since then the debate has run on two numbers that do not measure the same thing, and the regulator's own documents say so.

The number governments quote is accounts. eSafety's March 2026 compliance update, published under the social media minimum age framework, records that 4.7 million age-restricted accounts were removed, deactivated or restricted as at mid-December 2025, and that at the start of March 2026 over 300,000 additional age-restricted accounts were prevented from accessing platforms. The report itself immediately qualifies the figure: "the 4.7 million figure reflects the number of age-restricted accounts removed or restricted, not the number of users who have lost access to one or more age-restricted social media platforms", since individuals hold multiple accounts across multiple services, and some were inactive. It adds that figures for accounts actioned before or on 10 December 2025 "are not indicative of a platform's compliance" and "should not be used to make inferences about a platform's compliance".

The number the regulator commissioned measures behaviour, and it moved far less. Al Jazeera reported on 1 August 2026 that use of at least one age-restricted platform among surveyed children fell from nearly 86 per cent before the ban to above 81 per cent at the three month mark, with daily use moving from roughly 60 per cent to about 58 per cent. Roughly half of the children who kept accounts reported that the platform had never asked them to verify their age. Assistant Minister Andrew Leigh's response was that "we've had millions of accounts shut down" and that "we never expected that this would have 100 percent compliance".

The March compliance update supplies the mechanism, and it is documented rather than inferred. eSafety's pulse survey of 898 parents found that among parents whose child still had an account, the most common reason, given by 66.8 per cent, was that the platform had not yet asked the child to verify their age. Among parents whose child had lost an account, 43.6 per cent said the platform deactivated it, 36.3 per cent said the child did, and 26.6 per cent said a parent or carer did.

The report sets out four named compliance observations. First, messaging to under-16 users on some platforms encouraged them to attempt age assurance even where their declared age before 10 December 2025 was under 16. Second, some platforms allowed under-16 users to attempt the same age assurance method repeatedly until they obtained a 16 plus outcome. Third, reporting pathways for underage accounts have generally not been accessible or effective, particularly for parents. Fourth, some platforms appear not to have done enough to prevent children under 16 from having accounts at all.

eSafety focused on ten platforms it considers age-restricted: Facebook, Instagram, Kick, Reddit, Snapchat, Threads, TikTok, Twitch, X and YouTube. It has narrowed enforcement to five, Facebook, Instagram, Snapchat, TikTok and YouTube, and states it aims to decide on any enforcement action by the middle of 2026. It issued 23 legally enforceable information gathering notices and held 16 meetings with platforms. No penalty is recorded as issued.

What "reasonable steps" means remains legally undetermined, and the regulator says so plainly: the test is contextual, not prescriptive, and where court proceedings are pursued what is reasonable "is ultimately a question for the courts to determine on the evidence". On 25 March 2026 the Minister for Communications registered a new legislative rule tying the definition of an age-restricted platform more closely to features that pose the greatest risk of harm, which changes the perimeter of the scheme three and a half months after it started.

Parliament has since moved on penalties rather than on the definition problem. Al Jazeera reported on 27 June 2026 that the maximum penalty for systemic breaches would double from A$49.5 million to A$99 million, with the Prime Minister saying "it's clear Big Tech are not doing enough to comply with the law", and that eSafety would gain powers to demand documents from platforms, age-checking companies and app stores. The same report cited a peer reviewed evaluation in the British Medical Journal, surveying more than 400 children, that found insufficient evidence the ban significantly reduced social media use and noted substantial circumvention.

Meta published its own account on 12 August 2026. It says more than 750,000 accounts were removed across Facebook and Instagram by 30 June 2026, following more than 500,000 before the law took effect, that it uses AI detection of contextual signals and has raised its app store age rating to 16 plus in Australia, and that an education campaign in June and July 2026 reached about 1.3 million Australian users. Those are the company's figures about itself. Meta's policy argument is that the industry needs "a single reliable age signal delivered at the operating system or app store level".

Not known: how much of the residual 81 per cent has migrated to services outside the amended Rules, and whether any enforcement action will follow.