According to Reuters findings, Meta emits over 15 billion so-called risky ads — those that the system automatically flags as potentially fraudulent. In 2024, they brought the company over $7 billion in profit.
The problem is that in order for the ad to be removed from broadcast, the system must be in place 95 percent sure it's a scam. If the ratio is lower – the ad continues to show and its cost to the advertiser increases, which means more profit for Meta. This solution, as Reuters notes, is not the only one rewards riskbut also allows fraudsters to balance on the edge of acceptabilitybefore the system responds.
Moreover, an increase in the cost of a campaign is a warning signal to dishonest advertisers that they are “targeted” by the algorithm. Thanks to this, they can quickly modify content and bypass the detection system.
Read also: Meta accused of downloading adult videos. They were supposed to train AI models
Meta “cleans” ads where it is afraid of regulators
The documents show that Meta limits the scale of anti-fraud activities to markets where there is a risk of legal consequences. In countries with less active regulators, the company remains passive.
The report also shows that in the first half of 2025, Meta's fraud verification team could not take actions that would cost the company more than 0.15 percent. its revenues – so approx. $135 million. For comparison, the company's revenues during this period amounted to 90 billion dollars.
As Reuters notes, this means Meta treats the fight against fraud as an image costwhich must remain within acceptable limits for investors.
Read also: Meta is phasing out Messenger. Users need to prepare
The scale of fraud
According to Meta data, in 2025, 134 million ads were removedthat cheated or scammed users. The company also boasts that the number of reports from Internet users has decreased by 58 percent, but experts point out that this is still just the tip of the iceberg.
Meta contributes to one in three successful online frauds in the US, Reuters reports.
The data regarding the company's reaction to reports is also disturbing: 96 percent user reports of fraud were ignored or incorrectly rejectedand the scammers' ad accounts were able to operate even after hundreds of warningsif they have previously spent large sums on campaigns.
Read also: Meta disables political ads in the EU. Silence from Mark Zuckerberg
Meta has a budget for penalties
Meta, as the documents show, has set aside $1 billion for potential fineswhich may result from the lack of an effective fight against fraud. However, with annual revenue from risky advertising reaching seven times morefor the company it's simple cost of running a business.
As Reuters summarizes, Meta has created a system that formally fights fraud, but actually monetizes it — reacting only where regulatory or media pressure may damage the image.
