In a nutshell: Attackers systematically exploit AI branding in social engineering campaigns to manipulate employees — the attack vector is shifting from technical to behavioral vulnerabilities.
Cybercriminals are systematically adapting social engineering campaigns to the growing use of AI tools in enterprises. Microsoft and Google are documenting in parallel how attackers mimic AI platforms such as ChatGPT, Claude, or DeepSeek to distribute malware and steal credentials.
Microsoft Threat Intelligence warns that threat actors are leveraging global interest in AI itself as a social engineering lure. They impersonate ChatGPT, Microsoft Copilot, DeepSeek, or Anthropic’s Claude to distribute malware, steal login credentials, and commit financial fraud. Google documents in its latest Fraud & Scams Advisory a parallel evolution from classic phishing to Adversary-in-the-Middle attacks (AITM) and QR code phishing, as well as the growing abuse of trusted cloud services and AI-driven investment fraud.
Both warnings show that attackers are adapting established social engineering techniques to the growing role of AI in enterprise workflows, rather than relying solely on technical exploits. Microsoft documents, for example, ChatGPT-themed subscription renewal emails and counterfeit DeepSeek-V4 repositories with stolen trademarks distributing Vidar stealer malware. The campaigns continue to work through proven tactics such as urgency messaging, abuse of trusted services, and multi-stage redirect chains. Google estimates that global fraud losses in 2025 could reach approximately 580 billion U.S. dollars.
Both warnings underscore a paradigm shift: the attack vector is shifting from software stacks to the cognitive-behavioral level — to what employees believe, what they click on, and how they act when prompted by an AI-branded experience. According to IDC research, the majority of surveyed enterprises rank AI-enhanced phishing and impersonations, including deepfakes and voice clones, as a primary risk (58 percent). Analysts at Everest Group warn that the “Shadow AI” problem is not merely a visibility issue, but rather a trust problem: AI capabilities are increasingly emerging outside traditional IT controls and creating new attack surfaces.
Source: www.csoonline.com · Published June 9, 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 EU AI Act. Paraphrase and classification via Lumi News Pipeline v1.6.5.