LATIDIA · Investigación

Gobernar a la velocidad de la máquina: una arquitectura de inteligencia adaptativa para la aplicación de políticas de IA en tiempo real

arXiv: 2609.13466v1Tipo de anuncio: nuevo Resumen: La adopción de IA empresarial ha llegado al 78% de las organizaciones a nivel mundial, pero la infraestructura para gobernar esa adopción no ha seguido el ritmo. Este documento identifica y caracteriza

La noticia

arXiv:2609.13466v1 Announce Type: new Abstract: Enterprise AI adoption has reached 78% of organizations globally, yet the infrastructure to govern that adoption has not kept pace. This paper identifies and characterizes the attestation deficit, a structural condition in which organizations maintain governance policies but cannot produce auditable, tamper-evident evidence of enforcement within regulatory timelines. Drawing on empirical data from the Stanford 2026 AI Index Report (362 documented incidents), the IBM/Ponemon 2026 Cost of a Data Breach study (USD 4.99M average cost, 92% lacking access controls), and the EY/AIUC-1 Consortium survey (38% end-to-end monitoring, 17% agent-to-agent coverage), this paper demonstrates that the governance failure is organizational and architectural rather than technical. To