kAIte is an AI system that enables and represents multilevel citizen participation. A large number of citizens are reached through different ways and modes of collection to gather their input, depending on the type of project.

Based on keyword categorization, the AI system combines and differentiates all inputs while maintaining nuance. kAIte takes its place as a new stakeholder in local city meetings, bringing in the voices of carefully compiled personas of citizens, and being able to objectively talk on behalf of everyone.

Project duration

February 2023 – June 2023


Interactive Technology Design


Betsie Loeffen
Juwe van Vliet
Robin Smits


Centre for Urban Science & Policy

assessing coach

Tomasz Jaskiewicz

final grade

9,2 / 10

kAIte enables all citizens to influence urban planning actions and local governmental policies

Governing parties such as urban planners and governments often speak about citizens instead of with citizens. A crucial first step towards a more social and just society is asking citizens about their opinion, dreams and concerns, or what they are in need of to achieve their goals.

Through prototyping we learned that kAIte should not be integrated in current ways of working of governing parties, but rather react to what is being said. Therefore, kAIte’s tone of voice might be disruptive at times, reminding the attendees of a meeting that it is the citizens for whom they decide. To give kAIte the same hierarchical power as the other attendees, it has the same height as all other.

Citizen participation should take place on a large and equal scale

The public debate serves policymakers and citizens with a loud voice very well, while the largest group is often unheard. Thus citizens should be able to participate in various different ways, making it accessible for all. With all voices being heard.

Participating residents usually do not have a clear ‘persona’ of themselves, with clear one-sided needs and dreams, to be directly of use in a city council meeting. An AI might be able to make that translation. Through enacting the AI system ourselves, we found that such a system could label on the following aspects:

  • stakeholder profile with linked interaction intensity
  • quantity of stakeholders
  • topic or keywords mentioned
  • value judgement

Attendees of local city meetings should be in dialogue with hard and soft input from citizens

Citizen input in the form of demographic factual information is currently the ‘language’ that dominates. This kind of data is a limited translation of reality because it tells nothing about people’s values, beliefs and current or desired habits; all that is below the line of visibility (see image). As cities behave like large living organisms, being complex and networked, empathic tensions should also be present in local city meetings.