While Meta cannot offer services like Facebook and Instagram in China, Chinese advertisers are a significant pillar of its revenue. However, this relationship appears to conceal a huge ethical controversy.
According to Reuters NewsIn-depth investigation reportThe article reveals that Meta tolerated a large amount of advertising revenue from China for a long time in order to protect its huge advertising revenue.Advertising for scams, illegal gambling, and contraband originating from ChinaInternal documents reveal that even though Meta understood that these ads harmed users worldwide, under pressure to grow revenue, top management seemed to choose to "turn a blind eye," and even allowed fraudulent ads to resurface through policy shifts.
A pie worth $180 billion, with $30 billion in "toxic" revenue.
The report indicates that Meta's advertising revenue in China exceeded $180 billion in 2024, accounting for more than 10% of its global revenue. Meta estimates that approximately 19% of this huge sum (more than $30 billion) came from advertisements promoting fraudulent, illegal gambling, pornography, or other prohibited content.
This highlights the huge contradiction facing Meta: on the one hand, it needs to maintain platform security and user trust, but on the other hand, it is difficult to give up this "shady" revenue of up to several billion dollars.
Complex agent networks and "whitelist" vulnerabilities
Unlike other markets, Meta adopted a unique "agency system" in China, recruiting smaller "second-tier" agents through 11 "top-tier" resellers (large advertising agencies). This multi-layered structure created a perfect firewall, making it difficult for Meta to directly scrutinize the advertisers at the bottom.
Fraud groups use the following methods to evade detection:
• Identity concealment:Using fake accounts or stolen accounts.
• AI forgery:Use AI tools to generate fake verification documents.
• Technological cloaking:Purchase software tools to make the review system see a normal page, but when users click on it, they are redirected to a fraudulent website.
Even more serious is Meta's internal "whitelisting" mechanism. To avoid offending major clients, ads purchased through top Chinese agencies enjoy special protection: even if flagged as violations by the AI system, these ads are not immediately removed but remain online awaiting manual review. Due to insufficient manpower, the review process often drags on for several days or even fails altogether, giving fraudulent ads ample time to prey on victims.
Did Mark Zuckberg order the disbandment of the anti-fraud team? Meta spokesperson clarifies.
The investigation further revealed Meta's inconsistent attitude towards combating fraud:
• Early 2024:Internal sources have confirmed that China is the source of 25% of fraudulent advertising globally.
• Mid-2024 (Sweeping Period):By establishing an anti-fraud team and employing aggressive tactics, they successfully reduced the proportion of revenue from fraudulent advertising in China from 19% to 9%.
• End of 2024 (Turning Point):CEO Mark Zuckerberg reportedly intervened and demanded a "shift," subsequently disbanding the anti-fraud team and lifting the freeze on new advertising agencies, all in an effort to "unlock" revenue.
• Mid-2025:With the easing of regulations, the proportion of prohibited advertisements quickly rebounded to 16%.
In response, Meta spokesperson Andy Stone stated that the anti-fraud team was originally temporary, and Mark Zuckberg's directive was to "double down" on combating fraud globally, not to abandon related practices. The company claims to have blocked 4600 million illegal advertisements from its Chinese partners over the past 18 months.
Analysis: The eternal tug-of-war between safety and profit
In my opinion, this report once again starkly reveals the trade-offs that tech giants make between "shareholder interests" and "social responsibility." For Meta, the Chinese market is a very special case: it doesn't need to provide services there (there are no user maintenance costs), yet it can continuously collect advertising revenue from it.
This "take the money, don't take responsibility" structural advantage makes it difficult for Meta to have the incentive to eradicate fraud, since every blocked fraudulent ad represents a real loss of revenue. Although Meta emphasizes that it has been using AI to combat sophisticated criminal groups, until the design of "whitelists"—where business logic overrides security mechanisms—is removed, users' dynamic walls will likely remain difficult to clean up.
