Interesting Engineering has been covering applied science and engineering for 15 years. Our Facebook page has 16 million followers. Our editorial standards are verifiable: factually sourced articles and no political slant. By every measure Meta says it uses to define a legitimate publisher, Interesting Engineering qualifies.

Its Facebook monetization was cut anyway.

The demonetization arrived without a detailed explanation, without a cited violation, and, based on what IE’s team subsequently encountered, without any meaningful path to human review. Whether the decision was made entirely through automated processes remains unclear.

IE is not alone. Since mid-2025, creators across Facebook have reported sudden, unexplained monetization restrictions, the majority attributed to flags for “unoriginal content.” The pattern, which is abrupt enforcement, opaque reasoning, failed appeals, was formally documented on June 4, 2026, when Meta’s independent Oversight Board published findings that described Meta’s account governance as lacking “transparency and consistency,” producing removals that are not just unexplained but, in many cases, unexplainable.

How the system was built

The enforcement environment IE now operates in was constructed in stages, each one rational in isolation, collectively producing outcomes that some creators and publishers argue have affected legitimate accounts.

The first stage began in July 2025, when Meta announced comprehensive measures targeting accounts that repeatedly share unoriginal content without meaningful enhancement, building on enforcement actions that had already been taken against 500,000 accounts in the first half of the year. By November 2025, Meta had launched a dedicated content protection tool that automatically scanned Facebook and Instagram for Reels sharing similar traits with a creator’s original uploads.

The second stage came in March 2026, when Meta announced it was rolling out more advanced systems to handle content enforcement while reducing its reliance on third-party vendors. These systems were designed to take over repetitive, high-volume moderation tasks including flagging impersonation accounts, scam attempts, and harmful content, while human reviewers were concentrated on complex decisions and formal appeals.

The third stage was the March 2026 redefinition of “original content” itself. Meta updated its content guidelines to specify that content involving minor edits to another creator’s post — re-uploading posts the page had no role in creating, adding borders, inserting captions, or changing playback speed — would be classified as unoriginal, deprioritized in recommendations, and potentially demonetized if the behavior continued.

Taken together, the changes created a system with a narrower definition of eligible content and broader enforcement capabilities. Those systems are also retrospective; duplicate-content detection is designed to identify existing copies across the platform, not only catch new uploads in real time. As a result, archived content can become subject to enforcement long after it was originally published.

Why Meta automated

A 2025 Reuters investigation, drawing on internal Meta documents spanning 2021 to 2025, found that Meta had internally projected it would earn roughly $16 billion, approximately 10 percent of its annual revenue, from ads promoting scams and banned goods, including illegal gambling, banned medical products, and fraudulent investment schemes. The company was serving users roughly 15 billion higher-risk scam advertisements on average each day.

Meta’s policy was to ban advertisers only once its automated systems reached at least 95 percent certainty of fraud. Advertisers below that threshold but still flagged as higher-risk were charged premium rates rather than removed. One internal document from February 2025 specified how much revenue Meta was willing to forgo to tackle suspicious advertisers: 0.15 percent of total revenue, or $135 million, against what internal estimates described as a much larger problem.

Meta disputed the Reuters findings, saying the documents presented a selective view that distorted the company’s approach to fraud and that the 10 percent figure was a rough estimate, with subsequent audits finding many flagged advertisements were not actually in violation.

The Reuters findings help explain why enforcement became a strategic priority for Meta. Faced with persistent fraud, scam advertising, and growing scrutiny, the company increasingly relied on large-scale automated systems to identify and act against problematic accounts.

Critics argue that the unintended consequence has been a system optimized for enforcement at scale that can, at times, struggle to distinguish between bad actors and legitimate publishers.

What happens when it gets it wrong

The Oversight Board’s June 4 ruling was triggered by a specific case: an Instagram account with more than 70,000 followers, permanently disabled in 2025 for posting visual threats of violence against a journalist. The Board upheld that ban. But its investigation into the case opened a wider examination of how Meta handles account enforcement in general.

More than 750 people submitted public comments. Many described systems that simply did not function: no explanation for why their account was disabled, no way to appeal, and no ability to download their content. Many reported that decisions appeared to have been made automatically, with little evidence of meaningful human review, even on appeals involving longstanding accounts with large followings.

The frustration extends well beyond the Oversight Board’s comment process. By March 2026, nearly 60,000 users had signed a Change.org petition accusing Meta of “a widespread, systemic failure that has erased businesses, destroyed livelihoods, and cut people off from years of memories, relationships and vital communications.” One affected user, whose appeal was denied within minutes, said: “The appeal process was clearly not done by a person.”

CBS News, which contacted Meta about 35 affected accounts, found that at least 10 were later reinstated, meaning the original decisions were wrong, with none of the users ever told what triggered the ban in the first place.

The Board found that Meta’s account removals are not only unexplained but frequently difficult to challenge. Users whose accounts are disabled often encounter appeal processes that are opaque, impractical, or unavailable. Even paid Meta Verified support, marketed as providing direct access to a human agent, had failed to meaningfully assist some users who were wrongly actioned.

IE’s situation follows a similar pattern. A publisher with a 15-year record, 16 million followers, and no documented violations finds itself demonetized, with no specific violation cited and no clear path to obtaining a detailed review of the decision.

The payout numbers obscure the distribution

In 2025, Facebook paid content creators nearly $3 billion through its monetization programs; a 35 percent increase from the previous year and the highest annual total in the platform’s history. Meta has cited this figure prominently, and it is accurate.

It also says little about how those payouts are distributed. Sixty percent of total payouts went to Reels, with the remaining 40 percent distributed across Stories, photos, and text posts. The number of creators earning above $10,000 annually grew by more than 30 percent year-on-year.

Aggregate payouts rising while a specific segment of accounts loses monetization eligibility is not necessarily a contradiction. It reflects the possibility that gains are concentrated among certain creators while others experience sudden declines in reach or revenue.

The Oversight Board called on Meta to provide users with a dashboard showing current account status, past violations, and appeal options, with clear notifications at the time any penalty is imposed specifying the rule broken, the sanction applied, and how to challenge it. It recommended that Meta disclose whether and how automated systems were used in enforcement decisions. It also called for more effective appeals processes and for human review to be prioritized in edge cases.

Meta’s public response welcomed the ruling, stated it would review the Board’s recommendations, and committed to updating its position within 60 days. No specific commitments have been made.

The Board’s recommendations are advisory, and Meta is not required to implement them, though the company has historically acted on roughly 75 percent of Board recommendations, which makes its response to this particular ruling worth watching.

Disclosure

This article was reported and published by Interesting Engineering. Our Facebook account was recently demonetized. We have not received a specific violation notice and have been unable to obtain a detailed explanation for the decision. As a result, our publication is directly affected by the issues discussed in this article.

This article draws on publicly available reporting, Meta statements, Oversight Board findings, and our own experience navigating Facebook’s enforcement and appeals processes. We will update this article if Meta provides additional information regarding our case.

Our reporting remains available at interestingengineering.com.

Recommended Articles