随著生成式人工智慧的发展，骇客正利用深伪技术和高度客制化的网路钓鱼活动来发动攻击，例如北韩骇客组织假冒高阶主管进行视讯通话。这项技术不仅降低了攻击成本，还能透过操纵信任和紧迫感等人类情绪，使社交工程攻击变得更加容易且具说服力。

人工智慧同时也将整个诈骗供应链产业化，促使结合电子邮件、简讯和语音通话的多面向攻击日益增加，甚至能创造出完全虚构的合成身分。即时视讯通话已不再是验证身分的绝对保证，而大量逼真的伪造内容更可能导致人们对企业真实通讯的信任瓦解。

为应对这些威胁，企业开始部署由人工智慧驱动的进阶活体检测系统，透过分析生物讯号和背景数据来辨识真伪。此外，组织必须实施零信任架构，为敏感操作建立预先设定的通关密语，加强员工培训，并制定主动防御控制措施，以减轻无可避免的攻击所造成的损害。

With the advancement of generative AI, hackers are utilizing deepfake technology and highly customized phishing campaigns to launch attacks, such as a North Korean hacking group impersonating executives on video calls. This technology not only lowers the cost of attacks but also makes social engineering more accessible and convincing by manipulating human emotions like trust and urgency.

Artificial intelligence is also industrializing the entire fraud supply chain, driving an increase in multi-vector attacks that combine emails, text messages, and voice calls, and even creating entirely fictitious synthetic identities. Live video calls are no longer an absolute guarantee of identity verification, and the sheer volume of highly realistic fake content risks eroding public trust in genuine corporate communications.

To combat these threats, companies are deploying advanced, AI-driven liveness detection systems that analyze biological signals and contextual data to verify authenticity. Additionally, organizations must implement zero-trust frameworks, establish pre-agreed passphrases for sensitive operations, enhance employee training, and develop proactive defense controls to mitigate the damage caused by inevitable attacks.