Generative AI accelerates social engineering: convincing emails, voice cloning (vishing) and "CEO fraud" with deepfake audio and video. This particularly increases the risk of payment fraud and credential harvesting.
Typical signals
- Strange urgency around payments or changing bank account details
- Requests that try to bypass policy/four-eyes principle ("just quickly now")
- Voice/video calls with "off" timing, limited interaction, script-like behavior
Current-state check
- Process controls win: out-of-band verification + call-back procedures
- Train on deepfake scenarios; protect executives (contact hygiene)
- Monitor "lookalike" domains + brand impersonation
