Combating Vishing & Audio Deepfakes
Derick C.
July 08, 2026 8 min read
Vishing (voice phishing) has grown exponentially in complexity due to the availability of generative AI voice cloning software. It takes less than 3 seconds of reference audio to clone a voice with near-perfect accuracy. Organizations must update their defensive training to protect employees from advanced telephone scams.
How AI-Powered Voice Cloning Works
Attackers scrape public webinars, social media videos, or news interviews of corporate executives to harvest audio samples. They then use text-to-speech tools to impersonate CFOs, CEOs, or IT admins, calling junior staff to request urgent wire transfers, credential releases, or MFA bypass confirmations.
Developing a Verification Policy
To combat voice impersonation, companies should establish a strict verification protocol for out-of-band requests. No sensitive operation should be authorized solely on a phone call. Require multi-step approval through official chat channels or secondary managers.