What we said vs. what 2026 is showing
80-90% of social engineering uses Al voice, video or synthetic identity Human verification alone unreliable 30% won’t trust identity checks alone
62% of orgs hit by a deepfake attack Vishing surged 442% in 2025 Al scams up 1,210% in 2025 “Deception is outperforming detection”
62% of organizations suffered a deepfake social-engineering attack
442% surge in AI voice-phishing (vishing) during 2025
54% Al-phishing click rate vs 12% traditional (4.5x)
30% of enterprises will distrust standalone identity verification
Validated and accelerating. Al-generated voice and video deception is now mainstream and routinely outpaces human detection. Standalone identity verification is failing; assurance must move to multi-signal, provenance-backed verification with liveness and out-of-band checks. Train people to assume voice and video can be faked, and verify high-risk requests through a second channel.
2019-2021 Voice and video fraud rare, manual, and low quality
In 2024 – 85% of enterprises report at least one deepfake-enabledincident
Click-through rate (Al-generated phishing)
Click-through rate (Traditional phishing)
more effective than non-Al phishing
will use Al-generated voice, video, or synthetic dentities
Human verification alone becomes unreliable
50% of identity attacks to use real-time impersonation
30% of enterprises won’t trust identity verification alone
Deception is outperforming detection
