Meta Culpa

When a private platform can gag anyone without warning, explanation or meaningful appeal

Late one night, Meta informed me that I had breached its rules on cybersecurity.

This was surprising.

I had not hacked the Pentagon. I had not distributed malware, taught readers how to steal passwords or established a phishing operation from inside an unusually sophisticated pickle jar.

I had written an article about Donald Trump’s war with Iran back in March.

The post was three months old. It linked to a Substack essay containing political thoughts, history and commentary on a major international conflict. Meta’s systems nevertheless appears to have concluded that this represented some form of cybersecurity threat.

I appealed.

The post was restored.

That ought to have resolved the matter. Meta had reviewed the decision and apparently accepted that an article about geopolitics was not, in fact, malicious code wearing a necktie.

Instead, my account remained restricted.

The alleged offence disappeared. The punishment stayed behind.

No warning had preceded it. There had been no escalating sequence of notices explaining that the page was approaching some threshold. There was no coherent account of which sentence, link or image had supposedly breached the rule. There was no explanation of why content restored on appeal could continue producing penalties against the account that published it.

The machinery had spoken.

The machinery then wandered away without completing the paperwork.

That experience is personal, but it is hardly unique. Across Facebook and Instagram, creators, community organisations, small businesses and ordinary users regularly find themselves locked out, demoted, restricted or stripped of functionality by systems they cannot meaningfully interrogate.

They are told to consult Account Status.

Account Status tells them a restriction exists.

They are directed to appeal.

The appeal either returns a generic response or resolves one small component of the problem while leaving the broader penalty untouched.

They contact support.

Support sends them back to Account Status.

Eventually, they are informed that “the system” has made the decision.

This phrase deserves more attention than it receives.

“The system” is not a weather event. It did not drift over the Pacific and unexpectedly settle above Victoria Park. It was designed, commissioned and deployed by a corporation with a market value larger than the economies of many countries.

Yet once its decisions become inconvenient, the system is discussed as though it possesses independent constitutional authority.

A human representative may sympathise.

A support agent may agree that the situation appears strange.

Someone at Meta may even promise to escalate the issue.

But the system has the final say.

It is the digital equivalent of being arrested by a filing cabinet.

---

The Ministry of Community Standards

Meta describes its Community Standards as rules designed to protect safety and free expression across its platforms. It also provides users with an Account Status function showing some violations and restrictions, and offers appeals for many enforcement decisions. Eligible content decisions can sometimes be taken to the independent Oversight Board after Meta’s internal process has been exhausted.

On paper, this sounds like a functioning administrative structure.

There are rules.

There are decisions.

There are appeals.

There is even a board.

What is often missing is the connective tissue ordinarily expected in any system exercising serious power over people: clear reasons, consistent penalties, usable communication and a genuinely independent path for ordinary disputes.

The problem is not that Meta moderates content.

Any platform operating at its scale must do so. Facebook and Instagram contain scams, threats, abuse, violent material, coordinated manipulation and organised attempts to exploit vulnerable users. A platform that refused to intervene would quickly become a landfill with birthday reminders.

The problem is that Meta combines enormous enforcement power with deeply inconsistent procedural fairness.

It can restrict a page’s reach, remove monetisation, block posting, limit recommendations or disable an account. Those actions can affect income, professional reputation and access to an audience built over many years.

Yet the explanation may consist of a category label and a button.

“Cybersecurity.”

“Spam.”

“Account integrity.”

“Community Standards.”

These are not explanations. They are drawers in a very large filing cabinet.

A person trying to understand what actually happened is frequently sent through a chain of automated menus, help pages and support conversations that leads back to the original notice. Every doorway opens into the same room.

The enquiry does not progress.

It circulates.

Meta has constructed something resembling bureaucracy while removing the bureaucrat.

---

When the appeal succeeds but the punishment survives

My own recent case contains the central absurdity.

The post was restored.

The restriction remained.

That means one of two things must be true.

Either the post was wrongly removed and the associated penalty should have been reversed, or Meta believes the post was restored despite still justifying punishment against the account.

Neither possibility was explained.

This is not a trivial technical issue.

In any credible enforcement system, there should be a relationship between the offence, the evidence, the finding and the penalty.

If the finding is overturned, the penalty should ordinarily fall with it.

A court does not acquit the defendant and retain the prison sentence because the database has already become emotionally invested.

Meta’s system appears capable of separating content-level decisions from account-level consequences. A post may return while the account continues carrying the strike, restriction or loss of distribution associated with its original removal.

From the user’s perspective, the platform has admitted the prosecution failed but kept the conviction for administrative convenience.

The Oversight Board has repeatedly raised concerns about Meta’s systems of strikes, penalties and account governance. In June 2026, while upholding the permanent disabling of an account that had issued severe threats of violence, the Board still identified broader due-process concerns about how Meta disables accounts and governs account-level punishment.

A platform may be correct in one serious case while still operating a defective process overall.

Indeed, the most dangerous systems are not those that are wrong every time. Those collapse quickly. The durable ones are correct often enough to maintain authority while remaining opaque enough to resist accountability.

---

The endless loop

Anyone who has tried to resolve a platform problem knows the rhythm.

First comes the notification.

It tells you something has happened, though rarely enough to understand why.

Then comes the help page.

The help page explains the general policy in polished language, like a brochure produced by a country where nobody has ever experienced the actual government.

Then comes the appeal button.

Sometimes the appeal succeeds. Sometimes it fails. Sometimes it vanishes. Sometimes the decision is described as final without identifying who made it, what evidence was considered or whether anyone has read the disputed material.

If further support is available, the conversation often becomes more surreal.

A representative confirms that they understand the concern.

They say they can see the restriction.

They advise that the matter has been escalated.

Days later, another representative asks for the same information.

A third explains that the first department cannot assist because a different department controls the decision.

The different department cannot be contacted directly.

Eventually, everyone arrives at the sacred phrase:

The system has made the decision.

At that point, responsibility evaporates.

The support worker did not make the decision.

The policy team cannot review the individual case.

The review team cannot explain the account restriction.

The automated system cannot be questioned because it is not a person.

The company that built all of it remains somehow absent from the room.

This is corporate power perfected: every action has an owner until someone asks for accountability.

---

More than a website

Meta’s likely response would be that Facebook and Instagram are private services. People agree to the terms. The company owns the platforms and is entitled to enforce its standards.

Legally, that position carries considerable weight.

Socially and economically, it is no longer sufficient.

Meta is not merely operating a hobby forum where enthusiasts exchange model-train photographs.

Its platforms function as publishing infrastructure, advertising markets, community noticeboards, political arenas and commercial shopfronts. News organisations distribute reporting through them. Politicians address voters through them. Small businesses depend on them for customers. Charities organise supporters through them. Independent writers build audiences through them.

A restriction can therefore carry consequences far beyond inconvenience.

Imagine a shopping centre locking a retailer’s doors without notice, placing a sign over the entrance saying “security concern”, then refusing to identify the concern because the building-management system had reached a final decision.

Nobody would regard that as a satisfactory commercial process.

Yet digital businesses are expected to accept the equivalent because the landlord’s office has been replaced by a menu.

The scale of the platform changes the nature of the obligation.

The more economically and socially important a private system becomes, the less acceptable it is for enforcement to rely on unexplained automation and internal review alone.

That does not mean government should decide what Meta must publish.

It means people should have somewhere independent to go when Meta cannot coherently explain what it has done.

---

The missing referee

Australia already has bodies dealing with parts of the online environment.

The eSafety Commissioner can act in areas including serious online abuse, cyberbullying, image-based abuse and illegal or restricted content. It also accepts certain complaints about compliance with online-safety codes and standards. But it is not a general appeals tribunal for creators or businesses contesting an ordinary account restriction or wrongful content enforcement decision.

The Telecommunications Industry Ombudsman can resolve disputes involving telecommunications providers. People also contact it about digital-platform problems, but it currently lacks the jurisdiction to determine those complaints.

That gap is becoming increasingly difficult to ignore.

The TIO reported receiving 1,537 digital-platform complaints from 2023 to late 2025, describing a growing hole in consumer protection. By February 2026, that number had reached 1,780. The complaints included people trapped in automated recovery systems, unable to access accounts or obtain meaningful assistance. The TIO has explicitly called for its jurisdiction to be expanded into a broader Communications Ombudsman capable of resolving digital-platform disputes.

That proposal makes sense.

Australia needs a Digital Platforms Ombudsman, whether housed within an expanded TIO or created through another independent mechanism.

Not a Ministry of Truth.

Not an agency deciding whether political opinions are correct.

Not a tribunal through which every person irritated by a deleted meme can summon Mark Zuckerberg to Canberra.

A practical external dispute-resolution body.

---

The free-expression objection

Some people will hear the word “ombudsman” and immediately imagine government officials hovering over Facebook posts with red pens.

That concern is understandable.

Governments should not gain an easy mechanism to pressure platforms into suppressing lawful criticism or favouring official narratives. Political actors are not famous for acquiring powers and then displaying tasteful restraint.

But an ombudsman need not decide what people are allowed to say.

Its task would be procedural.

Did Meta follow its published rules?

Did it identify the alleged breach?

Was the penalty proportionate?

Was the appeal genuinely considered?

Did the company correct all consequences once the original decision was overturned?

Could the user reach a person capable of resolving the issue?

Those are questions of fairness, not ideology.

Banks, insurers, energy companies and telecommunications providers operate under complaint and external dispute-resolution frameworks. We do not regard that as government writing their customer correspondence.

Large organisations make mistakes, and internal complaint systems sometimes fail, and consumers need somewhere independent to go.

A social platform with billions of users should not enjoy a lower standard of accountability than a suburban electricity retailer.

---

Automation without accountability

Meta’s scale makes automation unavoidable.

No human workforce could manually assess every post, report, page and account across its services. Machine-learning systems are necessary to identify harmful material quickly and consistently.

But automation changes the nature of mistakes.

A human moderator may make one bad decision.

An automated classifier can reproduce the same misunderstanding thousands of times before lunch.

An article about cyberwarfare may be classified as cybercrime.

A historical image may be treated as extremist propaganda.

Satire may be read literally.

Journalism documenting a violent organisation may be mistaken for support.

A sudden surge in reports may resemble evidence of wrongdoing rather than an organised attempt to silence an account.

Automation brings speed and scale.

It also brings the capacity to be wrong with extraordinary confidence.

That is precisely why human review and external accountability become more important, not less.

Meta cannot deploy systems capable of affecting speech and livelihoods, then treat those systems as mysterious natural forces when challenged.

If the system has the final say, somebody must remain responsible for what it says.

---

My own modest cyber operation

I still do not know why an article about Trump and Iran was classified as a cybersecurity breach.

Perhaps the word “war” triggered something.

Perhaps the external link was incorrectly assessed.

Perhaps an automated system misunderstood the article.

Perhaps someone reported it maliciously.

Perhaps several systems interacted in a way that nobody inside Meta can now reconstruct.

I do not claim to know.

That uncertainty is the problem.

A company imposed an account-level restriction affecting my ability to publish and reach readers. It restored the supposedly offending post but provided no intelligible explanation for the original decision or the continuing penalty.

There was no identifiable decision-maker.

No coherent chain of reasoning.

No reliable path to independent review.

Only the system.

For a business built almost entirely upon communication, Meta has created a remarkable number of ways to avoid having a conversation.

---

Meta culpa

This is not ultimately an argument about my page.

Pages recover or disappear. Algorithms change. Audiences move. Publishers adapt because the internet has always been a continent undergoing active geological failure.

The larger issue is whether private platforms can continue exercising public-scale power through processes that would embarrass a suburban bowls committee.

Meta is entitled to establish rules.

It is entitled to remove genuinely harmful content.

It is entitled to act quickly against scams, threats and coordinated abuse.

But it should also be required to explain serious decisions, correct all associated penalties when those decisions are overturned, and submit unresolved disputes to an independent Australian body.

Power without reasons is arbitrary.

An appeal without access is theatre.

A support process that endlessly returns users to the decision they are challenging is not dispute resolution. It is a revolving door painted to resemble a corridor.

Australia does not need government deciding what Australians may say online.

It does need somewhere to go when a private corporation decides they may not say anything at all.

And when that corporation says the system has reached its final decision, someone outside the system should finally be allowed to ask why.

---

Thanks for reading, and more importantly, thank you for sticking with me.

There is a chance the page may be restored, and I am deeply grateful to the contact who has used his own network to escalate the issue within Meta. I am also fully aware that this is a privilege many creators, small businesses and ordinary users simply do not have.

Most people do not know someone who can reach inside the machine. They are left circling help pages, automated appeals and closed doors.

That is exactly why the wider issue matters.

Your support here, through subscribing, sharing and continuing to read, means more than ever.

Cheers.