LATIDIA · Ciberseguridad

Confundir el modelo, controlar el flujo: comprender y mitigar las fugas de privacidad de los agentes de LLM con el control del flujo de información

arXiv:2609.14003v1 Tipo de anuncio: nuevo Resumen: los agentes de IA personales basados en modelos de lenguaje grandes (LLM) tienen acceso cada vez más a los datos y comunicaciones privados de un usuario para proporcionar asistencia personalizada

La noticia

arXiv:2609.14003v1 Announce Type: new Abstract: Personal AI agents built on large language models (LLMs) are increasingly given access to a user's private data and communications in order to provide personalized assistance. This access creates a persistent privacy risk: the agent must decide whether a given sensitive information should be disclosed to a particular party. Existing defenses address this by making the agent's backend LLM more privacy-preserving through stronger system prompts, training, or explicit consent-checking procedures, but this approach has a structural challenge: whenever enforcement is a judgment the LLM makes over the same conversational context an adversary controls, the enforcement mechanism and the attack surface coincide. We demonstrate this against