[rwth-security] Deepfake Defense in the Age of AI