LATIDIA · Ciberseguridad

Combinar la memoria a largo plazo y los LLM tiene un alto coste: desde la inyección rápida hasta el compromiso persistente

arXiv:2610.28810v1 Tipo de anuncio: nuevo Resumen: La adopción generalizada de la inteligencia artificial generativa en los últimos años ha provocado un aumento paralelo en los ataques de inyección rápida y jailbreak. Al incrustar

La noticia

arXiv:2610.28810v1 Announce Type: new Abstract: The widespread adoption of generative artificial intelligence in recent years has brought a parallel rise in prompt injection and jailbreak attacks. By embedding malicious instructions, these exploits manipulate language models to bypass intended behavior and operational boundaries. In traditional chatbots, such attacks are confined to a single session because state memory is not preserved across conversations. However, persistent memory features in modern conversational interfaces enable cross-session attack propagation, establishing a threat vector known as memory injection. This work explores this emerging vulnerability in leading LLM-as-a-service platforms, highlighting its real-world severity and the need for countermeasures. While existing literature focuses on violations tailored to specific platform