Deepfakes Make Video Calls Dangerous—On-Device Detection Is The Answer
Deepfakes are undermining the trust historically placed in live video calls, enabling sophisticated social engineering attacks like the one experienced by Arup. While digital forensics can analyze past events, real-time detection is crucial to interrupt fraudulent activities during live interactions. Sending sensitive biometric data to cloud-based detection services introduces new privacy and security risks, making on-device detection a more secure and privacy-preserving alternative. AI
IMPACT On-device deepfake detection could restore trust in video communications and prevent AI-powered fraud.