Limitations in model generative accuracy
AI-generated deepfakes can create convincingly realistic but entirely fabricated information.
ENTITY
2 - AI
INTENT
3 - Other
TIMING
3 - Other
Risk ID
mit1077
Domain lineage
4. Malicious Actors & Misuse
4.1 > Disinformation, surveillance, and influence at scale
Mitigation strategy
- Integrate and mandate Multi-Factor Authentication (MFA) systems and deploy advanced, AI-powered deepfake detection platforms (e.g., Convolutional Neural Networks for artifact detection, liveness detection, and behavioral biometrics) to continuously analyze media and user behavior for signs of synthetic manipulation. - Establish comprehensive, ongoing media literacy and deepfake awareness training programs for all personnel to enable critical evaluation of content, and institute formal organizational protocols requiring cross-source verification before disseminating or acting upon sensitive or suspicious digital media. - Advocate for and adopt content authentication standards, such as the Coalition for Content Provenance and Authentication (C2PA) framework, including embedding digital watermarks and cryptographically-secured metadata at the point of media generation to enable clear traceability and signal content authenticity to end-users.