Meta Cut 750,000 Under-16 Accounts in Australia

For a company that runs its marketing through a Facebook page, the hardest risk to anticipate is not a competitor and not an algorithm change. It is a regulation that removes part of your audience from the platform on a given date.

In Australia, that has now happened. On 12 August 2026, Meta published its compliance record for the country's social media ban and disclosed that it had removed access to more than 750,000 Facebook and Instagram accounts it assessed as belonging to people under 16.

This is not a prediction that the Philippines will pass the same law. What the case does is show, with numbers attached, how platform-dependent marketing can shrink for reasons entirely unrelated to your own efforts. Below we read the development from one angle: how a Japanese-affiliated company in the Philippines should review its own acquisition design.

Part 1: Reading the News, Then Asking What It Means For Us

Start with the facts.

In a Newsroom post dated 12 August 2026, "Meta's Compliance with Australia's Social Media Ban," Meta published its enforcement record since beginning compliance in December 2025.

Here is what was disclosed, organised:

Meta named five measures behind those numbers:

Meta also stated its own position. Its primary concern is that platforms are not taking a consistent approach to how they remove teens, and that there should be a single reliable age signal delivered at the operating system or app store level. Age could be verified once, when a young person sets up a device or downloads an app, rather than repeatedly across the dozens of apps they use.

What is happening, then, is that a platform is removing its own users in order to comply with a regulation. And AI is doing part of the assessing.

Related: What Meta's 'Muse Image' Reversal Teaches About Generative AI, Likeness, and Personal-Data Risk explains this in detail.

Part 2: Key Terms Explained (For Decision Makers)

A few terms make this easier to judge.

What platform risk means When your acquisition depends on one platform, your results move with that platform's circumstances. Historically that meant algorithm changes and ad prices. It can now also mean users being removed for regulatory reasons. Nothing has to be wrong with your own activity for your reach to fall.

The idea of an age signal This is a design question about where, by whom, and how many times a user's age is verified. Meta argues it should happen once at the operating system or app store level rather than separately in each app. Whether that argument prevails determines who carries the implementation cost.

AI on the assessing side The assessment here is partly automated: age inferred from the context of posts and bios. This is not a story about using AI; it is a story about being assessed by it. Your own account and advertising are subject to the same class of automated judgement.

Fragmented compliance When regulation differs by country, platforms respond by country. The 16+ store rating is an Australia-only measure. The result is that the same service has different features and a different user base depending on where you are. For a company operating across borders, that is now the baseline condition.

Related: Avoiding Backlash from AI-Generated Social Posts: A Practical Guide for Japanese Companies in the Philippines explains this in detail.

Part 3: Applying This To Your Own Operation

For a Japanese-affiliated company in the Philippines, three things are worth working through.

The first is knowing where your acquisition actually depends.

Marketing in the Philippines tends to lean heavily on Facebook, for clear reasons: the user base is large, costs are low, and local staff are comfortable operating it. But concentration means the platform's circumstances become your results.

Three things are worth confirming now:

Produce the first row as a number. Not an impression — count the last three months of enquiries by channel. If it is above 80 per cent, that is dependency.

The second row is the substantive one. Followers on a platform are not your asset, because the platform can delete them. Contact details you hold are. Treat growing followers and obtaining contact details as two entirely separate activities.

The third row takes time. Traffic from search arrives months after you improve the content. That is precisely why it has to be started while dependency is still high and things are still working.

The second thing is preparing for automated assessment.

Meta states explicitly that it analyses posts, comments, bios and captions with AI. Age is the subject here, but the same mechanism serves other judgements. If your advertising or page is stopped by an automated assessment, there is no guarantee of reaching a human quickly.

The countermeasure is holding a route that does not depend on the assessment. Your own website, and the enquiries that come from it, are not subject to a platform's judgement.

The third thing is not over-predicting the regulatory outlook.

Whether the Philippines will pass a comparable law is not something I can tell you. What is knowable is that it happened in Australia, and that Meta has said platforms are not being consistent.

Rather than acting on a prediction, put yourself in a shape that is not badly affected whichever country's regulation arrives. That is the realistic form of preparation.

I helped a Japanese-run restaurant near Little Tokyo in Makati, with about 80 seats, adopt AI. We rebuilt an outdated website in WordPress, targeted small keywords with current SEO and GEO practice, and used AI in social media marketing centred on Facebook. We added an AI chatbot for simple enquiries and set the menu up so it could be rewritten easily from a PC. Consultation took about a week and the build one to two months.

We kept Facebook. It is still where the reach is in the Philippines. What we added was a route: from Facebook to the website, from the website to an enquiry. When everything closes inside Facebook, that day's reach becomes the ceiling on sales. With a route in place, the information left on the site keeps working in a month when reach falls.

Related: What OpenAI's 69-Page Report Does Not Show: There Is No Proof That Using More AI Makes You More Money | Case Study for Japanese Companies in the Philippines explains this in detail.

Part 4: Common Failure Patterns

Mistake 1: Concluding it is irrelevant because it is about Australia The regulation is indeed Australian. But everything this article recommends is useful whether or not a regulation arrives: counting your enquiry channels, building a route to your own site, holding contact details yourself. All of it pays off today.

Mistake 2: Jumping to "we should leave Facebook" Leaving Facebook in the Philippines is not currently rational. The user base is large and the cost efficiency is good. The problem is dependency, not use. Think in terms of moving from 80 per cent reliance to 60, not to zero.

Mistake 3: Treating follower count as an asset Followers are a number held on the platform's side. If the account is suspended, it is zero that day. The only figures that belong on your side of the ledger are the contact details you hold and the traffic to your own site. Make sure that distinction is shared internally.

Mistake 4: Leaving the alternative route until later Traffic from search arrives months after the content improves. Starting once the platform has a problem is starting too late. Beginning while dependency is high and there is still slack is the only approach that works.

Mistake 5: Assuming automated assessment behaves like human judgement Meta says supplementing report review with AI delivered higher accuracy and faster resolutions than human review alone. Decisions get faster; your opportunity to explain gets smaller. Hold a route outside the platform so that a stoppage is not a crisis.

Mistake 6: Arguing without numbers "We rely on Facebook too much" as a feeling will not produce a decision. Make one sheet counting the last three months of enquiries by channel. With numbers, the discussion moves on to how far to reduce it.

Three Practical Tips

Tip 1: Count your enquiry channels on one sheet Take the last three months of enquiries and split them into Facebook, search, referral and direct contact. It takes about an hour. Without that sheet, the dependency discussion stays at the level of impressions.

Tip 2: Build exactly one route to your own site From a social post to a specific page on your site, and from there to an enquiry. Build one. You do not need it on every post. One working route means information exists outside the platform.

Tip 3: Create a way to hold contact details yourself An enquiry form, a mailing list, a sign-up at the counter — the format does not matter. Have one mechanism that leaves contact details outside the platform. A hundred contacts you hold is worth more than a thousand followers you do not.

Bonus: How to Make Use of PH AI Works

At the stage of working out where your acquisition depends, an outside view sometimes helps. PH AI Works offers a 30-minute free consultation for Japanese-affiliated companies in the Philippines.

We will work through it against your actual operation: how to count enquiry channels, how to build the route to your own site, how to scope an AI chatbot if you are considering one, and where to start with search and GEO (optimising to be cited by generative AI). It is fine to talk at a stage where you have decided nothing.

Sources & References

• Meta Newsroom, "Meta's Compliance with Australia's Social Media Ban" (published 12 August 2026, updated 13 August 2026; over 750,000 accounts removed, reporting period 1 December 2025 to 30 June 2026, AI detection, campaign reaching approximately 1.3 million, Meta's position on age signals): https://about.fb.com/news/2026/08/metas-compliance-with-australias-social-media-ban/
• Meta Newsroom (official newsroom): https://about.fb.com/news/
• National Privacy Commission, "Data Privacy Act of 2012" (the Philippine framework where you hold contact details yourself): https://privacy.gov.ph/data-privacy-act/