Participatory AI development model

Artificial intelligence has rapidly become a part of everyday life in Finnish society, yet public trust in those who develop and use it remains low. In the development of generative AI, individuals’ sensitive data forms a significant resource, meaning that citizens should have genuine opportunities to influence the direction in which AI is developed and how it is used. This is particularly true in the public sector, where decisions have broad impacts on daily life.

Trust in AI is built on transparency, inclusion, and the recognition of citizens’ epistemic rights. The Finnish Constitution states that citizens must have the possibility to influence decisions that concern them. This right should also extend to the development and deployment of AI. Finns would like to influence decision making and shared societal matters easily and quickly online, but the discussion culture on social media discourages many from participating.

Without transparency and genuine agency, sustainable trust in AI cannot be established. Public administration and other societal institutions should view the safeguarding of citizens’ AI agency as a core responsibility. The framework of epistemic rights summarises citizens’ rights in the age of AI into four areas: access to information, access to skills and capabilities, participation, and the right to form opinions free from pressure. Strengthening these rights is essential for maintaining trust and legitimacy.

The model for participatory AI development and the technical citizens’ AI rulebook – developed through bold experimentation – are examples of how citizens’ perspectives can be incorporated into AI development and decision-making.

The citizen dialogue What do you think about AI, Finland? organised by Sitra and Reaktor, was carried out in September 2025 on the anonymous, constructive discussion platform Voxit and engaged nearly 6,500 participants. The discussion published 194 statements about AI development and use, of which 169 were created by citizens.

The aim of the dialogue was to address challenges of trust from the citizens’ perspective and to create a rulebook for the use of AI in public services. The discussion sought to increase public understanding of AI, give citizens agency, and test and refine the rules created by citizens. The goal was to produce a technical rulebook for guiding a language model and to document the entire operating model in a way that can be utilised elsewhere in Europe.

The dialogue produced a rich and diverse dataset of citizens’ views on the use of AI in public services. Based on this material, a technical rulebook was created to influence the operation of AI, and it is presented in this memorandum. The result is a valuable foundation that supports the responsible and democratic use of AI in society and helps to identify solutions for safeguarding public trust.