Our latest threat report examines how malicious actors combine AI models with websites and social platforms—and what it means for detection and defense.