MODA: Meta's High Market Share a Target; Urges AI Adoption to Reduce Fraud

Taiwan's Anti-Fraud Command Center has identified Meta's social media platforms as the primary channel for scams, demanding algorithm changes. The Ministry of Digital Affairs (MODA) stated on the 21st that due to Meta's high market share, it is easily exploited by fraudsters. The government will work with Meta to introduce AI to fight AI-driven scams and reduce their reach. From January to April this year, Meta-related fraud cases accounted for over 84% of the total, with financial losses exceeding NT$1.36 billion.
政策NQ 70/100出典:PR Times

📋 Article Processing Timeline

  • 📰 Published: May 21, 2026 at 15:00
  • 🔍 Collected: May 21, 2026 at 15:31 (31 min after Published)
  • 🤖 AI Analyzed: May 21, 2026 at 15:35 (3 min after Collected)
(CNA, Taipei, May 21, by reporters Lai Yu-chen and Kao Hua-chien) The Executive Yuan's Anti-Fraud Command Center stated yesterday that social media platforms owned by Meta are the primary channel for fraud groups, demanding that the company modify its algorithms to reduce the reach of scams. The Ministry of Digital Affairs (MODA) said today that Meta's high market share makes it susceptible to exploitation by fraud groups, and the government will move forward with Meta, introducing AI to combat AI in an effort to reduce scam exposure.
The Anti-Fraud Command Center announced yesterday that according to statistics, from January 1 to April 30 this year, there were 14,073 fraud cases involving Meta's social media platforms, accounting for 84.31% of all cases. The total financial loss amounted to NT$1,361,781,574, representing 86.2% of the total loss. This shows that Meta's platforms remain the main channel for fraud groups, with agricultural product shopping scams on Threads being a major issue. Meta must immediately and proactively modify its algorithms and detection systems to block repetitive fraudulent posts at the source; otherwise, the government will, if necessary, amend laws to introduce corresponding control measures and penalties.
MODA Deputy Minister Hou Yi-hsiu stated at a post-Executive Yuan press conference today that Meta has already joined the 'anti-fraud good guys team,' proactively taking down 7.8 million fraudulent posts last year and over 20,000 fraudulent accounts this year. However, anti-fraud efforts can always be more aggressive. Therefore, the government continues to communicate with Meta on various methods, hoping to reduce the public's exposure to scams through the technologies Meta possesses.
Hou said that current source management solutions include strengthening verification and management of high-risk accounts, safety reminders for external links, and AI customer service. MODA will also continue to communicate with Meta regarding the algorithm adjustments pointed out by the Anti-Fraud Command Center. According to MODA data, Meta's Facebook platform was fined four times by the ministry last year, with total penalties amounting to NT$18.5 million.
Lin Chun-hsiu, Director-General of MODA's Administration for Digital Industries, added at the press conference that the government acts according to the law and will issue fines where necessary. He noted that Meta's cooperation in fraud prevention is still high, but its large market share makes it a target for fraudsters. The government will work with Meta on technological and legal fronts, using AI to fight AI, to reduce the reach of scams. (Editor: Su Chih-tsung)

FAQ

為何台灣政府將Meta平台視為主要的詐騙管道?

根據台灣打詐指揮中心的統計,今年1月1日至4月30日,Meta所屬社群平台涉及的詐騙案件數佔總數的84.31%,財損金額佔比高達86.2%,總金額超過13.6億新台幣。

台灣政府提出了什麼解決方案來應對Meta平台的詐騙問題?

政府要求Meta修改演算法與偵測系統,並將與其合作,導入「以AI對抗AI」的策略。具體方案包括強化高風險帳號驗證、使用者導外連結安全提醒及AI客服等。

Meta對於防堵詐騙採取了哪些措施?

根據數發部次長侯宜秀的說法,Meta已加入政府的「打詐好人隊」,去年主動下架780萬則詐騙貼文,今年也下架了2萬多個詐騙帳號。

如果Meta不配合修改演算法,政府有何後續手段?

打詐指揮中心表示,若Meta不立即主動修改演算法,必要時將修法提出相對應的管制措施及罰則。數發部去年已對Facebook平台裁罰4次,合計1850萬元。

哪一類型的詐騙在Meta平台上特別嚴重?

報導中特別提到,在Meta旗下的Threads平台上,農產品購物詐騙為大宗。