読書リサーチレーダー投資フレームワーク
ログイン / 新規登録
← 読解に戻る
ニューストピック · 思科AI钓鱼攻击重大2 分で読めるDisconfirmAI 編集部

Cisco Talos: Chinese Hackers UAT-11985 Suspected of Using AI to Generate Phishing Emails Targeting Taiwan Research Institutions

Cisco Talos disclosed an APT campaign targeting Taiwan think tanks, suspected of using AI to generate highly similar phishing emails and tampering with event poster QR codes.

2026-10-19 11:54 GMT+8

Cisco's threat intelligence team Talos disclosed on October 19, 2026, a targeted spear-phishing campaign specifically aimed at research institutions in Taiwan. The activity is attributed to the Chinese hacking group UAT-11985, characterized by the use of suspected AI-generated content to create highly convincing social engineering lures, combined with Adversary-in-the-Middle (AiTM) techniques to steal authentication credentials.

Talos noted that UAT-11985 sent phishing emails impersonating organizations such as the 'Taiwan EU Center', the 'Center for International Relations Studies' at National Chengchi University, and the 'Taiwan Thinktank'. These emails share extremely similar structures and wording patterns: they first establish a professional image with lengthy geopolitical and policy analysis, then add personalized remarks like compliments, exclusive invitations, or VIP seats for recipients, and finally provide event time, location, and registration details. Due to the close similarity in sentence patterns across different emails and the ability to quickly swap out event themes and recipient information, Talos assesses that the attackers likely used AI-assisted generation for these phishing contents to achieve low-cost mass customization.

In terms of propagation methods, the attackers employed a hybrid 'digital + physical' strategy. Beyond email, UAT-11985 embedded poster images obtained from event websites into their phishing letters but maliciously tampered with the QR Codes on the posters to point to attacker-controlled servers. Talos believes the attackers clearly expected recipients to print these posters and display them on office bulletin boards, thereby spreading phishing links from personal inboxes to public physical spaces and significantly expanding the potential reach of victims.

ログイン / 新規登録
読書リサーチレーダー投資フレームワーク

On the technical front, the campaign deployed an advanced AiTM phishing framework. After clicking the link, victims are first presented with a fake Google Form webpage, followed by a forged Google login interface. The tool transmits account, password, and verification data via HTTP POST and maintains real-time connections using WebSocket, allowing attackers to dynamically switch between password input, passkey, and multi-factor authentication (MFA) screens based on Google's actual verification state, thereby bypassing traditional static anti-phishing defenses.

ソース

  • ithome.com.tw↗
トピック

思科AI钓鱼攻击

思科威胁情报团队(Cisco Talos)披露,追踪代号UAT-11985的高级持续性威胁(APT)组织疑似利用人工智能辅助生成钓鱼邮件,针对台湾研究机构发动攻击。该组织通过仿冒台湾欧盟中心、政治大学国际关系研究中心等机构名义,结合篡改线下…

5 ステップ1 件更新日 2026-10-19 18:45

経緯

  1. 2026-06-25UAT-11985发起的钓鱼邮件中提及的活动日期之一(据第三方分析推断),攻击者利用真实公共活动信息作为掩护。pk-sharma.com ↗
  2. 2026-10-08思科Talos发布详细分析报告,披露UAT-11985利用AI辅助生成钓鱼邮件及AitM框架攻击台湾研究机构。blog.talosintelligence.com ↗
  3. 2026-10-12CyberPress报道该攻击细节,强调攻击者通过HTTP POST和WebSocket混合架构实时同步认证状态,以捕获会话令牌。cyberpress.org ↗

未解決の問題

  • 是否有更多受害机构确认收到此类钓鱼邮件?
  • UAT-11985背后的具体国家支持实体身份是否已被其他情报机构证实?
  • 针对此类AitM攻击,现有的反钓鱼措施(如FIDO2密钥)在实际部署中的有效性如何?
トピック全文を読む →
リサーチを続ける →