Executing Red Team Scenarios: Proactive Adversary Emulation with Machine Learning
Erdener Uyan and Gökberk Gülgün present an open-source machine learning framework for generating dynamic, automated red team attack scenarios.
Summary
The Complexity of Scripting Realistic Attack Scenarios
Manual red team engagements require extensive operational planning to orchestrate believable multi-stage attack paths mimicking specific APT behavior.
Designing a Proactive ML Attack Generation Framework
Erdener Uyan and Gökberk Gülgün introduce an open-source tool that uses machine learning algorithms to generate realistic decision trees for adversary simulation.
Automating Multi-Stage Lateral Movement and Execution
The authors demonstrate how the framework selects optimal exploit techniques and persistence mechanisms dynamically based on discovered host telemetry.
Evaluating Blue Team Detection Coverage and Resilience
The presentation discusses measuring detection efficacy, identifying coverage gaps, and continuously training defensive response teams with automated scenarios.
Related talks
Red Team Village·1 hr
Bruce Schneier · Anita Nikolich · Chris Cottrell · Rich Harang · Ram Shankar · Omar Santos
Bruce Schneier, Omar Santos, and industry AI leaders debate the future of artificial intelligence in offensive security, automated exploitation, and red teaming.
Red Team Village·1 hr 36 min
Bruce Schneier
Security technologist Bruce Schneier shares perspectives on AI-driven hacking, automated vulnerability discovery, and the societal implications of machine-speed attacks.
Blue Team Village·1 hr 5 min
Olaf Hartong · Mauricio Velazco
Presents open-source tooling and methodologies to measure defensive coverage and test detection engineering pipelines.