Deepfake Detection
Methods that detect synthetic or manipulated media (audio, images, video) used to impersonate people in identity verification and biometric systems.
Overview
Deepfake detection aims to identify AI-generated or heavily manipulated media intended to defeat identity verification (e.g., selfie checks) or to enable fraud via impersonation.
How it’s used in identity systems
- As a signal alongside PAD/liveness, device checks, and document verification.
- To flag suspicious submissions for step-up or manual review.
- To harden enrollment and re-verification, where synthetic media can create or take over accounts.
Common challenges
- Rapidly evolving generation methods and attack techniques.
- Domain shifts (lighting, cameras, compression) that can affect detector performance.
References
Vendors using Deepfake Detection
Latest Data Cards
Data Card Incode says Deepsight is first to pass iBeta Level 3 PAD on both iOS and Android
2026-03-02CC-BY-4.0paddeepfake-detectionibetaIncode said its Deepsight face-liveness platform passed iBeta's ISO/IEC 30107-3 Level 3 presentation-attack test on both iOS and Android, positioning the result as a dual-platform first.
- The result was certified by iBeta on February 16, 2026 under ISO/IEC 30107-3, the main independent PAD test framework used in commercial procurement.
- ID Tech reported that Incode claimed a zero-percent error rate across 900 Level 3 presentation attacks spanning both mobile platforms.
- The claim reflects rising market pressure for third-party validation against sophisticated spoofing and deepfake-driven attack scenarios.
Data Card Mitek enhances video verification to counter injection attacks in Spanish market
2026-02-16CC-BY-4.0video-injection-detectiondeepfake-detectionmitekMitek enhanced its SEPBLAC-compliant digital onboarding platform for Spain with four defense layers against injection attacks—deepfake detection, digital manipulation analysis, injection attack protection, and face gallery analysis—achieving 99.9% accuracy on known deepfake engines and over 99% detection rates for face swaps.
- 89% of Spanish companies reported increased fraud attempts annually, with identity fraud the leading cause of corporate fraud loss in Spain.
- The update aligns with February 2025's new European injection attack standard, CEN/TS 18099:2025.
- Mitek's approach combines AI-trained detection with traditional software checks targeting coordinated fraud attempts during onboarding.