OWASP Atlanta Chapter

Coordinated adversarial information operations often span multiple platforms, but detection systems struggle with limited labels, changing tactics, multilingual content, and platform-specific behavior. This talk presents lessons from building a cross-platform detection pipeline that learns from unlabeled Telegram activity and applies those signals to detecting campaigns on X.
Using temporal coordination, behavioral patterns, content signals, and account relationships, we examine which signals transferred across platforms, where automated detection broke down, and why campaign-level attribution remains difficult.
Attendees will leave with a clearer understanding of how these signals can be combined, which signals generalize across platforms, where machine learning systems require human oversight, and how detection teams can build more robust workflows as adversarial behavior changes.