Towards Measuring and Mitigating Social Engineering Software Download Attacks
summary
Researchers deploy a DPI collector on a large academic network for two months, reconstruct 36,000 executable downloads and identify ~2,000 successful social‑engineering (SE) downloads; ads start ~81% of chains, ads dominate and repackaged + entice is 48.5% of cases. AV detects <5–25% at day‑zero and >70% by day‑30. They extract ad‑chain and download‑URL features (domain age, Alexa rank), train a realtime classifier that achieves ~91% true positive and ~0.5% false positive.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 2:10Research Background
- 4:18Categorization
- 12:10Measurement
- 18:39Detecting
- 22:28Conclusion
related talks
30mSecurity Starts With You: Social Engineering
55mSocial Engineering Overview-How To Prevent Your Co. from Getting Hacked
53mAI-equipped Threat Actors Versus AI-enhanced Cyber Tools: Who Wins?
44mHacking Humans: Social Engineering Techniques
50mHacking Humans : Social Engineering Techniques and How to Protect Against Them
58mSocial Engineering: Hacking Humans
35mProtecting the Organization: The Psychology of Social Engineering
41mHacking Humans : Social Engineering Techniques and How to Protect Against Them