An interactive demo showcasing DIANA, a novel architecture for ethical AI agents that structurally separates language generation from normative reasoning, ensuring safer and more auditable AI decision-making.