Urban planning in a democratic context faces a persistent tension: decisions about the long-term shape of a city require input from the people who live in it, but the tools available for that input are rarely up to the task. Traditional consultation methods — online surveys, public hearings, phone calls — struggle to reach diverse populations, tend to attract the same engaged minorities, and seldom inspire genuine creativity around complex, long-term questions like climate adaptation, housing, or mobility.

Aarhus faced exactly this challenge. The municipality needed a way to engage citizens on issues that felt abstract and distant — futures that were hard to visualize, tradeoffs that were difficult to articulate — while ensuring that the process remained genuinely open, not just symbolically participatory. The question was whether digital tools could expand who gets to participate, and whether AI could help citizens express ideas they would otherwise struggle to put into words.

The response was crea.visions, a Hybrid Intelligence platform that combines generative AI with participatory design workflows. Rather than presenting citizens with predefined options to react to, the platform invites them to co-create: generating AI-assisted visualizations of preferred futures, annotating them with ideas and concerns, and linking them to specific locations in the city. The output is not a survey result — it is a living map of citizen imagination, grounded in real places and real values.

For Aarhus, the platform was tailored in close collaboration with municipal officials and local stakeholders. The design choices reflected a core Hybrid Intelligence principle: AI should lower the barrier to participation, not substitute for the human judgment that gives participation its meaning. Facilitation and moderation remained central throughout — trained facilitators guided participants through the process, ensuring that the AI-generated outputs sparked dialogue rather than short-circuiting it.

This design reflects what the underlying research framework calls Hybrid Intelligence Foresight (HIF): a methodology that integrates generative AI with collective intelligence to enhance participatory foresight at scale. The crea.visions platform has been deployed across 11 interventions in Denmark, France, Italy, and Serbia, with over 7,000 participants co-creating more than 30,000 futures artefacts — evidence that the approach can travel across institutional and cultural contexts without losing its participatory character (Sherson et al., 2025).

What made the Aarhus deployment distinctive was its integration into the municipality’s actual planning processes. Citizen-generated visions were not collected and shelved — they fed into deliberative forums and concrete planning discussions, creating a direct line between public imagination and institutional decision-making.

Over 300 citizens participated in the Aarhus pilot, generating thousands of unique visions tied to specific locations across the city. The project demonstrated that hybrid intelligence — citizen-AI co-creation — can meaningfully expand who participates in urban planning, and what quality of input they are able to contribute. Participants who might have found a traditional survey alienating found the visual, generative format accessible and engaging.

Aarhus is now exploring how to embed the tool into long-term civic engagement strategies — moving from pilot to infrastructure. The ambition is not just to run better consultations, but to build a sustained capacity for participatory foresight: a city that can engage its citizens in genuine futures thinking, not just reactive feedback.

For municipalities navigating the same tension between democratic ambition and practical constraint, Aarhus offers an early model: use AI to lower the barrier to participation, keep human facilitation at the center, and design the process so that what citizens imagine actually reaches the people making decisions.

Reference: Sherson, J. et al. (2025). A Theory of Hybrid Intelligence Foresight: Evidence from GenAI-Powered Large-Scale Community Engagement. Preprint available at ResearchGate: https://www.researchgate.net/publication/397379763