Intelligence and AI at the Convergence Zone of Data and Threats
summary
Forecasts tie three concrete trends: cheap hyperscale data, labeled threat intel, and foundation-model AI. Microsoft reports dwell time falling from ~200 to ~20 days, introduces a nation-state/cybercrime taxonomy and names like Mint Sandstorm (Iranian subgroups using same-day POC exploits and custom Drokbk/Soldier malware). A criminal toolkit, Storm-1101, sells AiTM (Adversary-in-the-Middle) proxies for MFA bypass; cracked Cobalt Strike kits prompt a multi‑party takedown effort that successfully disrupts distribution.
AI-written summary · about
chapters (12)
- 0:00RSAConference 2023
- 5:07A good forecast requires data and analysis
- 8:40Three stacked forces
- 10:32A clearer taxonomy for understanding threats
- 14:56Actor Overview
- 20:33Mint Sandstorm
- 22:48Overview of AiTM
- 28:32Actors known to use Cobalt Strike
- 30:56Digital Crimes Unit action
- 33:44Artificial intelligence
- 35:21Al and Security Copilot
- 45:02Strength in numbers
related talks
01h 19mAI vs AI: Cyber Defence and Managed Resilience
11mAI-Powered Cybersecurity in the Age of Advanced Threats
44mAI Meets Cybersecurity: Boosting Speed Without Losing Control
27m"Cybertelligence": How AI is changing the face of cybersecurity
37mFundamentals of AI safety and security
57mAI as the New Hacker Frontier
40mAI As The New Hacker Frontier
43mThe challenge of AI in cloud security