Meta was found to have derived 10% of its revenue from fraudulent advertising, and some have suggested it was deliberately delaying its response to the fraudulent ads.
Reuters reports, based on internal documents obtained exclusively by the company, that Meta is making huge amounts of money from fraudulent advertising, and that these profits are supporting huge investments in fields such as AI, but that it is not keeping up with the reality that it is doing enough to prevent this.
Meta is earning a fortune on a deluge of fraudulent ads, documents show | Reuters
Bombshell report exposes how Meta relied on scam ad profits to fund AI - Ars Technica
https://arstechnica.com/tech-policy/2025/11/bombshell-report-exposes-how-meta-relied-on-scam-ad-profits-to-fund-ai/
Meta estimates that it earns 10% of its revenue from scams, report says | TechCrunch
https://techcrunch.com/2025/11/06/meta-estimates-that-it-earns-10-of-its-revenue-from-scams-report-says/
In its internal forecast for the end of 2024, Meta estimated that $16 billion, or approximately 10% of its total annual revenue in 2024, would come from ads for fraudulent and prohibited products. According to an internal document published in December 2024, Meta's platform displays an estimated 15 billion 'high-risk ads' to users every day, which are clear indicators of fraud. Another document from late 2024 stated that this category of ads alone generates approximately $7 billion in annual revenue.
According to Reuters, Meta has a unique approach to advertisers it suspects of fraud. Rather than immediately banning advertisers if the likelihood of fraud is less than 95%, Meta applies a mechanism called a 'penalty bid.' This involves charging suspicious advertisers a higher advertising rate as a penalty, with the aim of discouraging fraudsters.
However, this means that Meta will sell fewer fraudulent ads, but will make more money from the ones it does sell, thereby partially offsetting the revenue loss. Furthermore, it has been pointed out that Meta's ad personalization system means that once a user clicks on a fraudulent ad, they are more likely to see more fraudulent ads.
There have been reports of fraudulent advertising, such as when a recruiter for the Royal Canadian Air Force had his Facebook account hacked and used to post cryptocurrency scams. Despite the recruiter and his colleagues reporting the scam more than 100 times to Meta, the response was delayed, resulting in an acquaintance of the recruiter being defrauded of 40,000 Canadian dollars (approximately 4.36 million yen).
The reason why Meta has been reluctant to take measures against fraudulent ads is likely due to its huge investment in AI development. The company plans to spend up to $72 billion (approximately 11 trillion yen) on AI-related capital expenditures by 2025, and relies on its advertising business for the funding.
In the past, Meta reportedly restricted its safety staff's use of computing resources in order to focus resources on AI and virtual reality (VR) in 2023. A 2022 document also noted that Meta classified fraudulent advertising as a 'low-severity' problem and was not actively investing in automated detection.
In fact, a 2023 document estimated that of the approximately 100,000 legitimate fraud reports submitted by users per week, 96% were ignored or incorrectly rejected. Regarding fraudulent advertising, the US Securities and Exchange Commission (SEC) is investigating Meta for financial fraud, and UK regulators also cited Meta's services as being involved in 54% of payments fraud losses in 2023.
Meta internally predicted that it could face fines of up to $1 billion from regulators, but another document estimated that revenue from 'high-risk fraudulent advertising' would reach $3.5 billion in just six months, far exceeding the fine. The company also expressed concern that a sharp decline in fraudulent advertising revenue would impact its business forecasts. A February 2025 document revealed that its team responsible for reviewing suspicious advertisers was not authorized to take any action that would result in a loss of more than $135 million, or 0.15% of the company's total revenue.
According to Reuters, when the Singapore Police reported 146 cases of fraud in the past, Meta staff deemed 77% of them to be non-violative, saying that they violated the spirit of the policy but not the letter.
In response to Reuters' report, Meta spokesman Andy Stone countered that the documents 'present a selective view that distorts Meta's approach to fraud.' He also claimed that the 10% revenue forecast was 'broad and overly inclusive' and had been subsequently revised downward, but did not specify the revised figure. Stone also highlighted the company's success, saying, 'Over the past 18 months, user reports of fraudulent advertising have decreased by 58% globally, and we will have removed more than 134 million pieces of fraudulent advertising content by 2025.'
However, an internal review of Meta in April 2025 concluded that 'Meta's platform is more prone to fraudulent ads than Google's,' suggesting that the company may be lagging behind its competitors in its fraud prevention efforts, Reuters reported.
Related Posts:
in Web Service, Posted by log1i_yk