Přeskočit na obsah

Data Privacy - R3V3RS3D Lecture Summary

This comprehensive presentation explores the intersection of data privacy, cognitive warfare, and open-source intelligence (OSINT) in the digital age. The lecture examines how personal data is generated, collected, and weaponized in modern information operations.


🌁 Odkaz na přednášku

  • Personal security checklists and best practices
  • Adversarial techniques (makeup, clothing patterns) to evade facial recognition
  • WiFi mapping and Bluetooth beacon tracking
  • Digital footprint analysis across platforms (Google, Facebook, social networks)

Resources:


  • Data Brokers: Major players include Acxiom, Oracle Data Cloud, Experian, Equifax, TransUnion
  • What They Collect: Demographics, financial data, online behavior, health information, location history, political affiliations
  • Who Buys It: Marketers, employers, insurance companies, law enforcement, political campaigns
  • Regulation Status (2025): Patchwork laws in US (CCPA/CPRA); GDPR restrictions in EU

The presentation details how psychological operations exploit human cognitive vulnerabilities:

Core Concepts:

  • Overton Window Shift: Normalizing extreme ideas through gradual exposure
  • Memes: Cultural units that spread ideas, emotions, and tribal identities
  • Narratives: Story frameworks that promote specific worldviews
  • Dark UX Patterns: Manipulative design tactics including misinformation loops, echo chambers, fear appeals

Psychological Exploits:

  • Cognitive biases (confirmation bias, bandwagon effect)
  • Emotional decision-making over logic
  • Information overload leading to uncritical acceptance
  • Social identity manipulation (us vs. them tribalism)

Real-World Example: Operation SIMCARTEL - Europol operation that seized 40,000 SIM cards used to create 49 million fake online identities for fraud and influence operations.

Research Reference: Cognitive Warfare Academic Paper


Four AI/NLP Techniques for Narrative Intelligence:

  1. NER (Named Entity Recognition): Identifies persons, organizations, locations, dates
  2. POS (Part of Speech): Grammatical categorization for syntactic analysis
  3. Topic Modeling: Discovers hidden thematic structures using BERTopic
  4. Sentiment Analysis: Classifies emotions (polarity: -1 to +1; subjectivity: 0 to 1)

Applications:

  • Automated detection of disinformation campaigns
  • Narrative tracking across multiple sources
  • Identifying coordinated inauthentic behavior
  • Protecting analyst mental health through automation

Tools & Research:


Ethical Guidelines:

  • Use only legally accessible data
  • Respect privacy and Terms of Service
  • No doxxing, harassment, or vigilantism
  • Consider potential harm

Essential OSINT Tools & Resources:

Web Archives & Search:

Image Intelligence:

Geospatial:

Infrastructure Research:

Frameworks & Training:


GitHub Repository: UNICRI-Decode-CW Contains code samples, analysis notebooks, and methodology for cognitive warfare detection.


Jindrich Karasek Senior Cyber Threat Researcher, PSU CAS Linktree Profile