Deepfake

Deepfakes are AI-generated or manipulated media that convincingly imitate real people in face, voice, or video. In KYC and authentication they enable high-grade impersonation: synthetic selfies, lip-synced videos, cloned voices, or face swaps intended to defeat biometric checks. Risks include account opening fraud, account takeover, and social engineering of support staff.Countermeasures combine presentation attack detection, active video liveness, secure capture with device attestation and cryptographic nonces, and anomaly detection on temporal jitter and compression artifacts. Operationally, raise thresholds for high-risk events, diversify signals with MRZ or NFC checks and registry hits, and run adversarial tests. Education and agent scripts reduce help-desk vulnerability. Governance documents threat models, testing coverage, metrics like attack error rate, and rapid update paths as tools evolve.