Critical Weakness in AI - Attention Hijacking

Overview

Uncover critical vulnerabilities in multi-agent AI systems where Large Language Models can manipulate each other through persuasion dynamics and attention hijacking techniques.

---

Subjects

Computer Science