Mitigation
AML.M0034Deepfake Detection
What it is
Apply deepfake detection algorithms against any untrusted or user-provided data, especially in impactful applications such as biometric verification, to block generated content.
Detectors may use a combination of approaches, including:
- AI models trained to differentiate between real and deepfake content.
- Identifying common inconsistencies in deepfake content, such as unnatural facial movements, audio mismatches, or pixel-level artifacts.
- Biometrics analysis, such blinking, eye movements, and microexpressions.
References
Related by meaning· 6
Nearest entities by semantic similarity across the cs-graph corpus.