Meta tag

Sigurnost AI

2 articles

db#1125

Google DeepMind’s six AI traps: The web is a minefield

DeepMind’s new study turns the web into an adversarial playground, detailing six ways autonomous AI agents can be hijacked via everyday tools like APIs and documents.

01 Apr 2026

db#666

LLM safety gets a math upgrade—but will it outrun attacks?

ES2 weaponizes the geometry of embedding spaces to widen the gap between safe and toxic prompts, turning a structural flaw into a defense.

24 Mar 2026