Fabricated Realities: AI Personas and Simulacra and the Threats They Pose
summary
Voice cloning moves from minutes to 3 seconds of audio and enables realistic synthetic video and audio used in extortion and phishing at scale; provenance breaks down so attackers can fabricate logs and transcripts. Demonstrated mitigations include frequency filtering of published audio, asking interlocutors to perform a visible gesture (open fingers between face and camera — finger-gap video check), and cryptographic provenance / chain-of-trust for recordings to prove authentic sources.
AI-written summary · about
chapters (9)
- 0:00Introduction to AI personas
- 6:11Rise of synthetic identities
- 10:46Deepfakes and cybercrime surge
- 15:32Consequences of AI-generated deception
- 19:25Proving authenticity in digital age
- 22:20Manipulation techniques in media
- 25:23Ethical dilemmas of AI cloning
- 31:02Evolution of voice cloning technology
- 33:52Defending against AI impersonation
related talks
39mSynthetic Trust: Exploiting Biases at Scale
36mDemocratized Deception: How Threat Actors Are Weaponizing AI to Scale Cybercrime
39m Put Words In My Mouth
47mFighting AI with AI
30mAI-Generated Synthetic Media, aka Deepfakes and its impact
33mYour voice confirms my identity
30mReality Under Attack: Just How Serious Is the Deep Fake Threat?
45mFUD vs. Reality: Bracing for an AI-Powered Threat Landscape