Meta announced that, from March 2024 to December 2025, through collaboration with the Ministry of Digital Development, the Ministry of the Interior, and the Criminal Investigation Bureau in Taiwan, it successfully removed as many as 7.8 million messages targeting Taiwanese users.Fraudulent advertisements.
According to data from the Criminal Investigation Bureau's "165 Anti-Fraud Dashboard," this wave of technology-driven law enforcement has been remarkably effective. Since August 2024, both the number of fake investment fraud cases and the amount of financial losses have decreased by more than 70%.
Three key technological advantages: real-name registration for advertisements, AI detection, and facial recognition for celebrities.
Meta stated that this success was primarily due to its long-term implementation of three major anti-fraud strategies, including several key technologies specifically designed for the Taiwan market:
• Pioneer of "Advertiser Verification":This is a mechanism that Meta pioneered in Taiwan. By fully and mandating verification and locking down the identities of accounts advertising in Taiwan, the threshold for fraud groups to abuse the advertising system has been directly raised, making it much harder for them to simply apply for unverified accounts and run ads indiscriminately as before.
• Celebrity facial recognition technology:To combat deceptive advertisements impersonating Morris Chang, Jensen Huang, or financial commentators, Meta has implemented facial recognition technology. The system can proactively scan, identify, and remove harmful content that misuses celebrity images before the public reports it.
• AI-automated detection:Using AI to identify fraudulent signals can block fraud networks at their source.
Data shows that reports of fraudulent ads on global social media have decreased by more than 50% over the past 15 months, indicating that automated blocking has indeed been effective.
Advertising has proven difficult, so scam groups are turning to private messaging.
Although there are fewer advertisements, the fraud groups have not disappeared; they have simply moved to a different battlefield.
Meta has issued a special warning that as cross-border fraudulent ads are being blocked in large numbers, scam groups are shifting significantly to using non-advertising channels. With Christmas and New Year's Eve approaching and people exchanging greetings frequently, scam groups are starting to use the "impersonation of acquaintances" method to contact people via Messenger or private messages, tricking them into providing personal information or clicking on phishing links.
Meta urges the public to make good use of account security features (such as two-factor authentication) and to be wary of common traps such as unknown links, high-paying job offers, or online dating.
Analysis: This is a never-ending game of cat and mouse.
In my opinion, the data released by Meta this time truly reflects the platform's determination to govern, especially the "advertiser verification" measure, which has effectively increased the operating costs of fraud groups. I believe that when you browse Facebook or Instagram recently, you should see a slight decrease in the frequency of those poorly made celebrity fraud ads (although they still exist, they are not as exaggerated as before).
However, this battle is far from over. As the effectiveness of "casting a wide net" advertising diminishes, fraud groups will turn to more precise and covert "one-on-one" private messaging scams, and even use deepfake voice or video calls to impersonate your relatives and friends.
The government and platforms can build the first line of defense, but the last line of defense lies in the vigilance of users. The battle against AI will only intensify in the future, so when you receive messages online that claim you "need money urgently" or "have a great deal," be sure to take an extra second to verify them.
