Skip to content

ODSP2100 Pre-recorded On demand

Understand how attackers use AI to create and evade malware

Kieran Human

Summary

From Microsoft's AI summary

Kieran Human of ThreatLocker discussed how attackers use AI to create malware and evade detection, citing quarterly CVE counts rising from about 5,000 in 2021 to over 37,000. Demonstrations included an AI assistant performing ransomware-like steps, an LLM producing data-stealing code after rephrasing, Evilginx phishing that captured a session cookie despite MFA, a Bitcoin clipboard hijacker using a USB Rubber Ducky, and ClickFix. Recommended defences were application allow listing, deny-by-default policies, ringfencing and Zero Trust Cloud Access.

Session description

Microsoft's catalogue description

AI is changing how attackers create, evade detection, and deliver malware, while giving defenders new ways to identify threats. See demonstrations of AI-assisted malware creation, exploit detection, and delivery methods, and learn how to distinguish human-driven techniques from those enabled by AI.

← All sessions