Forty-one per cent of chief information security officers surveyed by Gartner reported at least one social engineering incident involving a deepfake during an employee audio call in the preceding 12 months. Thirty-six per cent reported an incident involving a deepfake during a video call.
Gartner surveyed 297 CISOs and equivalent senior cybersecurity leaders between March and May 2026. Respondents could report more than one type of incident. Email phishing, spear-phishing or business email compromise was reported by 79 per cent, while 58 per cent reported an incident involving voice or text-message phishing.
The figures describe incidents reported by the surveyed security leaders; they do not establish how many attacks succeeded or how often deepfakes are used across all organisations. Gartner says AI can make impersonation attempts harder to assess by combining synthetic audio or video with information gathered about a target.
Gartner advises organisations to verify sensitive requests through a trusted channel, regardless of whether they arrive by email, phone or video. It also recommends stronger controls for account recovery, privileged access and payment approval, alongside incident procedures that connect impersonation reports with unusual account or transaction activity.
