Skip to main content

AI Deepfakes and Biometric Security: A Conversation with Rohan Pinto

N2K NetworksJune 4, 202530 min398 views
26 connections·40 entities in this video→

Privacy Violations and Data Breaches

  • πŸ” Meta and Yandex tracking scripts are deanonymizing Android users by exploiting browser features to link web activity with app identities, bypassing security models.
  • ⚠️ A compliance automation firm, Vanta, disclosed a data breach due to a product code change that led to cross-customer data leakage, exposing employee information and security configurations.
  • πŸ€– A new Linux botnet named Pumabot is targeting vulnerable IoT devices, particularly surveillance systems, by brute-forcing SSH credentials for cryptocurrency mining.
  • πŸ“ˆ The Ramnit banking trojan has seen a significant spike in operational technology (OT) environments, likely repurposed to extract industrial control system credentials.
  • πŸ”’ The North Face experienced a credential stuffing attack, exposing customer personal details due to password reuse across different platforms.
  • πŸ’₯ An Indian grocery delivery startup, KiranaPro, suffered a devastating data wiping attack after hackers accessed root accounts, leading to the deletion of all its data and halting operations.

AI Deepfakes and Biometric Security

  • πŸ’‘ Rohan Pinto, CTO of 1Kosmos, discusses the increasing threat of AI deepfakes to biometric security, noting their ability to mimic facial features, voice patterns, and iris scans.
  • πŸ›‘οΈ Traditional biometric systems like Face ID are insufficient against deepfakes as they primarily identify presence rather than binding identity to the authentication mechanism.
  • 🧐 1Kosmos's Live ID approach verifies authenticity in real-time using forensic analysis, liveness checks, and matching against registered biometrics to ensure a live individual is present.
  • πŸ”¬ Advanced techniques include analyzing depth, iris position, shadows, infrared data, and detecting noise or image clarity issues to differentiate real individuals from masks or deepfakes.
  • 🀝 A multi-layered approach combining biometric factors, behavioral analytics (typing patterns, voice), and passive detection techniques enhances accuracy and reduces spoofing success rates.
  • βš™οΈ Organizations should enable liveness detection on both edge and server sides, incorporate behavioral analytics, and ensure data governance with standard-based technologies.

Emerging Cyber Threats and Defense

  • 🎯 The pro-Ukraine hacker group Black Owl is identified as a significant cyber threat to Russian institutions, utilizing custom tools and targeting government agencies.
  • 🚨 CISA has released critical advisories for vulnerabilities in Schneider Electric and Mitsubishi Electric industrial products, posing risks to critical infrastructure.
  • πŸ‡¬πŸ‡§ The UK has established the new Cyber and Electromagnetic (CyberEM) Command to integrate cyber operations and electromagnetic warfare into its military strategy.
  • πŸ’» Sophos researchers uncovered over 130 open-source GitHub projects booby-trapped with backdoors, targeting other hackers and indicating a potential 'distribution as a service' racket.
Knowledge graph40 entities Β· 26 connections

How they connect

An interactive map of every person, idea, and reference from this conversation. Hover to trace connections, click to explore.

Hover Β· drag to explore
40 entities
Chapters11 moments

Key Moments

Transcript109 segments

Full Transcript

Topics18 themes

What’s Discussed

MetaYandexPrivacy ViolationData BreachPumabotIoT DevicesRamnit MalwareICS IntrusionsCredential StuffingBlack OwlCISA AdvisoriesData Wiping AttackCyberEM CommandAI DeepfakesBiometric Security1KosmosLiveness DetectionGitHub Backdoors
Smart Objects40 Β· 26 links
CompaniesΒ· 16
ProductsΒ· 10
MediaΒ· 1
PeopleΒ· 3
ConceptsΒ· 7
EventsΒ· 3