Disrupting malicious uses of AI | February 2026

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

Source: OpenAI — Published February 25, 2026 at 00:00 UTC — Category: Models