A compliance story with two incompatible headlines
At an August 14, 2026 Senate inquiry hearing, Meta, Google/YouTube and TikTok told lawmakers exactly what they've been doing since Australia's under-16 social media ban took effect on December 10, 2025: removing accounts at scale. Meta has deactivated 756,000 Facebook and Instagram accounts suspected of belonging to under-16s, up from roughly 500,000 in March. Google blocked about 740,000 account holders from YouTube. TikTok removed 550,000 accounts when the rules landed and has been pulling roughly 24,000 more a month since (Reuters, via Investing.com).
Those are large numbers by any standard — eSafety separately confirmed more than 750,000 Meta accounts alone had lost access by June 30, 2026 (eSafety Commissioner). But eSafety Commissioner Julie Inman Grant told the same inquiry that roughly 80% of Australian minors are still accessing social media months after the ban took hold — a figure drawn from the regulator's own research (Al Jazeera). Both things are true simultaneously: platforms are removing hundreds of thousands of accounts, and children are recreating them or simply signing up on the platforms that never asked their age in the first place. Deactivation counts measure enforcement effort, not the underage population — a point the platforms themselves made to the inquiry, urging caution against reading early removal data as a proxy for the ban's actual effect.
The regulator's response: bigger stick, not better instruments
Rather than resolve that measurement gap, the government's answer is the Online Safety Amendment (Strengthening Enforcement for the Social Media Minimum Age) Bill 2026, passed by the House of Representatives on July 1, 2026 and now before the Senate (Parliament of Australia). It would double the maximum civil penalty for a systemic breach from A$49.5 million — the ceiling set by the original Online Safety Amendment (Social Media Minimum Age) Act 2024 — to A$99 million, and expand eSafety's power to compel documents, not just information, from platforms and third-party age-assurance vendors during investigations (ABC News).
The case for the expanded powers is genuinely strong, and worth taking on its own terms before dismissing it. Inman Grant told the inquiry that under current law her office can demand information but not primary documents — and without those documents, she said, eSafety would "likely be unable to discharge the onus of proof required in any civil penalty proceeding" (ABC News). That is not a rhetorical flourish: it is a regulator explaining, under oath, that the law as written may make its own enforcement power practically unusable. A regulator that cannot compel the evidence needed to prove a case is a regulator that cannot enforce the law Parliament actually passed. If eSafety loses in court on a technicality of evidence access rather than on the merits of whether a platform complied, that's a real deficiency worth fixing.
Where the fix outruns the diagnosis
But doubling the fine ceiling does nothing to close the gap between "accounts removed" and "children off the platform," and Inman Grant herself undercut the case for urgency: she told the inquiry the ban was "showing early signs of working," citing a modest decline in underage accounts even as she cited the same 80% figure that suggests otherwise. Digital Industry Group Inc, the platforms' local trade body, made the sharpest version of this point to the inquiry: expanded powers are being sought before a single fine has been issued under the existing regime, meaning lawmakers are legislating for a hypothetical evasion problem rather than a demonstrated one (ABC News).
The Law Council of Australia's submission raises a narrower but concrete risk: document-compulsion powers broad enough to reach "third parties" could, without tighter drafting, sweep in the children and parents the ban is meant to protect, rather than staying confined to platforms and age-assurance vendors (ABC News). Communications Minister Anika Wells has said the powers target vendors, not users — but that assurance is a policy intention, not a statutory limit, and the distinction matters once the bill has enforcement teeth.
The better fix is measurement, not multiplication
A$99 million is a bigger number than A$49.5 million, but it is not a better number if the underlying problem is that neither eSafety nor the platforms can reliably say how many under-16s are actually on the platforms at any given time — only how many accounts got flagged and removed. The Age Verification Providers Association's submission to the inquiry gets closer to the actual fix: mandatory, independent audits of how platforms are age-assuring users, rather than self-reported removal counts that Meta, Google and TikTok control and disclose on their own timetable (ABC News). Document-compulsion power that lets eSafety verify those self-reported numbers against platforms' actual detection systems is a legitimate and proportionate response to the evidentiary gap Inman Grant described. A fine ceiling twice the size of one that has never been triggered is a signal of seriousness, not a mechanism that closes the 80% gap. Parliament should pass the powers; it should treat the fine increase as symbolic until there's a compliance metric worth attaching real penalties to.