This recipe explores what it means to intentionally design an AI system to support democratic exchange. Rather than focusing only on technical performance or user satisfaction, it encourages you to consider how design choices shape participation, representation, and legitimacy.

It introduces a core design idea from the KT4D project: building in meaningful friction. That might sound counterintuitive, but the aim is to slow down interaction, make assumptions visible, and prompt users to question what the AI is doing and why. These friction points—like profiling questions, visible datasets, or collaborative prompting—help shift AI from a “solution engine” to a deliberation support tool.

This recipe draws directly from the design and development of the Digital Democracy Demonstrator, used in citizen-facing labs in Krakow and Madrid. It provides questions and design principles for those developing or procuring AI tools to be used in democratic settings.

This recipe is part of the Digital Democracy Lab Handbook — a practical resource for facilitators, technologists, and public sector actors exploring how AI can support meaningful democratic exchange.