Meta co-hosted the "National Anti-Fraud Action Forum" with the Taiwan Digital Development Department and the Criminal Police Bureau of the Ministry of the Interior's National Police Agency today (7/22), and joined hands with private organizations including the Taipei Computer Association, MyGoPen, and the Taiwan Digital Security Development Association to jointly promote anti-fraud cooperation plans. Meta also announced its three upgraded anti-fraud strategies in Taiwan, and announced that it has removed more than 2024 million fraudulent advertisements targeting Taiwanese users in Taiwan since 430. It also announced an investment of NT$100 million to assist in the development of AI anti-fraud tools to further strengthen the protection network.
Simon Milner, Vice President of Public Policy for Asia Pacific at Meta, said that as transnational fraud groups become more organized and their methods become more sophisticated, the removal of fraudulent ads and accounts must rely on automated technology and cross-border collaboration. In addition to strengthening platform policies, Meta also works closely with the Taiwanese government and private partners to actively block the spread of fraudulent information on social platforms.
Three strategic upgrades: comprehensive upgrade from technology to collaboration
This time, three major anti-fraud strategies were announced in Taiwan, including strengthening the platform's anti-fraud measures, clearing the network of fraud groups, and working with partners to build a defense line.
Among them, Meta will further enhance its ability to identify fraudulent content through the "ABC" three major frameworks (Actor, Behavior, Content). It currently requires all accounts that advertise in Taiwan to complete real-name authentication and clearly mark the information of investors and beneficiaries. In addition, it also uses facial recognition technology to actively block fake celebrity advertisements to prevent people from being deceived by images and other factors.
By combining artificial intelligence with behavioral models provided by security experts, Meta can more quickly locate and combat fake accounts across fraud zones such as Myanmar, Laos, the Philippines, and the Middle East. According to data released by Meta, more than 2024 million fraudulent accounts have been cleared worldwide since 700, thereby blocking fraudulent activities on service platforms such as Facebook.
In addition to cooperating with the government, Meta also participates in the Global Signal Exchange to promote cross-border information sharing. In Taiwan, it has been cooperating with the Criminal Police Bureau, the Taipei Computer Association, and the MyGoPen platform that provides fact-checking since 2021, through in-depth social platforms and physical activities, to further promote fraud awareness education in Taiwan.
According to the monitoring data of the Taiwan Digital Security Development Association, the probability of Taiwanese people encountering fraudulent advertisements on social platforms has decreased significantly recently, especially for fake investment and celebrity fraud content, with the number of reports decreasing by 90%. Deputy Minister of the Ministry of Digital Development Lin Yijing also emphasized at the forum that thanks to the passage of the Anti-Fraud Law by the Legislative Yuan, the communication between the government and the platforms has become closer, and the expectation for the platforms' proactive anti-fraud capabilities will continue to increase.
Interior Undersecretary Ma Shiyuan called for the public to prevent fraud as a long-term battle. It is not just a government task, but also requires the private platform to play its corporate responsibility, invest in technology, and create a trustworthy digital environment. The Criminal Police Bureau also pointed out that in the future, it will deepen cooperation with Meta and other platforms to promote the exchange of anti-fraud experience and technology upgrades, so that all people can become fraud-spotting gatekeepers.
Meta: Will not profit from fraudulent advertising
In a subsequent interview, Simon Milner pointed out that global fraud is actually roughly the same, including celebrity fraud, online shopping fraud, investment fraud, health and wellness fraud, love fraud, etc., and the channels of fraud have also changed from text messages and phone calls in the past to the widespread use of online social platforms. The reason is that it is easier to achieve the purpose of fraud by launching a psychological offensive through an account directly linked to an individual.
Since current fraud has the characteristics of low risk, low cost and high return, since the global epidemic in 2020, when many people's lives rely heavily on the Internet, many fraud activities have begun to increase, and even evolved into many new forms of fraud. The fraud methods will change in response to reports, investigations, regulations, or in conjunction with new technologies. For example, there are currently more and more frauds carried out through generative AI.
In addition to proposing three major anti-fraud strategies, Simon Milner also emphasized that Meta will not profit from fraudulent advertising, and that it will do its best to prevent such problems. Through cross-border and cross-unit cooperation, the ability of criminals to profit through fraud will be improved, the cost will increase, and their profit margins will continue to be compressed, thereby significantly reducing fraud in online communities.
Simon Milner said that the reason for choosing to promote the anti-fraud strategy in Taiwan and even further investing in MyGoPen to create an anti-fraud tool that allows people to directly compare whether the content is fraudulent is mainly because Taiwan has a very high Internet usage stickiness rate and a high smartphone usage penetration rate. Therefore, he hopes to work hand in hand with the Taiwan government and anti-fraud agencies to promote the public's anti-fraud awareness and further strengthen the ability to identify fraud, thereby reducing the chance of fraud.
