Overview

• Deepfake incidents now commonly pair synthetic audio or video with social engineering to impersonate executives and push employees into wiring funds or granting access.
• Gartner also reports 32% of organizations experienced attacks on AI applications in the past year, including prompt injection that attempts to steer model outputs.
• Vendors are beginning to embed deepfake detection into collaboration platforms like Microsoft Teams and Zoom, though large-scale effectiveness remains unproven.
• Security leaders are urged to require application-level approvals with phishing-resistant MFA so no single call or video can authorize high-risk actions.
• Some firms are running awareness simulations using executive deepfakes, while experts stress that detection signals are probabilistic and should be layered with stronger processes.