AI-powered service desk attacks leverage generative AI to produce polished emails, chat messages, and realistic voice or video impersonations, making social engineering more convincing. Attackers use AI to automate reconnaissance by scraping publicly available information to personalize their requests, increasing their credibility. AI also enables attackers to scale attacks by generating multiple phishing variants and adapting their tactics to bypass service desk defenses. These attacks commonly target onboarding processes where new employees are less known to IT staff, exploiting the urgency and routine nature of service desk requests. Effective prevention requires replacing assumed trust with verified identity, using technologies like biometric liveness detection and secure onboarding solutions that eliminate insecure credential delivery.

The impact includes increased risk of unauthorized access to organizational accounts and systems through social engineering at the service desk. Attackers can gain credentials, reset multifactor authentication, or approve sensitive changes by impersonating legitimate users, potentially leading to account compromise and subsequent data breaches or financial losses. The threat is particularly acute during onboarding, where attackers exploit gaps in identity verification. While no known exploits in the wild are reported, the enhanced credibility and scalability of AI-powered attacks raise the likelihood of successful compromises if mitigations are not implemented.

No official patch is applicable as this is a social engineering threat enhanced by AI rather than a software vulnerability. Organizations should implement stronger identity verification methods at the service desk, especially for onboarding and sensitive actions like password resets. Recommended mitigations include using biometric liveness detection to confirm the presence of a live person, avoiding sending credentials via insecure channels such as SMS or email by using secure enrollment links, and enforcing identity verification before approving sensitive requests. These measures reduce reliance on agent judgment and help defend against AI-powered impersonation attacks.