Apurahat ja residenssipaikat

Tiede

Citizen Agency in AI? Democratizing Algorithmic Services in the City

Hakemuksen tiivistelmä

Algorithmic tools are increasingly being incorporated into public sector services in cities today. The rapid growth of these emerging technologies and services necessitates new forms of algorithmic literacy and awareness among everyday citizens for sufficient understanding in grappling with their implications. How can we involve a wide range of citizen voices in bottom-up engagement, and increase algorithmic literacy and civic agency to make algorithmic public services more accessible? The project aims to understand citizens’ algorithmic literacy, agency and participation in the design and development of AI services in the Finnish public sector in order to advance more democratic and citizen-centric digital infrastructures. The project identifies the following research objectives: 1) understanding the values, narratives and discourses embedded in public sector data-centric and algorithmic services, 2) understanding citizens’ level of literacy and perceived agency with regards to algorithmic public services, 3) empowering citizens to critically engage with algorithmic public services, and 4) transforming design of public sector AI services to ensure civic participation. The role of language is amplified in the translation of information between these domains, both in shaping wider public discourses and in creating transparent and accessible technical documentation. Because the wider impact of algorithmic services, or AI systems in general, is more than the sum of their ones and zeros, extending beyond technical details, this project requires an interdisciplinary exercise of interpretation. The project will offer policy recommendations for Public Administrators and Digital Service Providers, and digital citizenship guides highlighting tools for Citizen Recipients and Citizen Advocates in exercising civic rights and co-designing imaginaries that truly democratize data and algorithmic services in the public sphere.

The CAAI (Civic Agency in AI) project examines citizens' algorithmic literacy, agency, and participation in the design and development of AI services in the Finnish public sector. The goal is to advance more democratic and citizen-centric digital infrastructures at a time when public institutions are increasingly turning to AI-powered services, and when the stakes of getting this right are high.

A central part of the project examined how AI is talked about and governed. Using a mixed-methods approach combining Natural Language Processing with qualitative analysis, the project compiled a digital corpus of AI-related texts from the EU, spanning policy documents, stakeholder feedback on the EU AI Act, industry communications, and news media. This open-access dataset reveals how diverse actors, such as policymakers, industry, and researchers, frame the risks and benefits of AI, and how contested those framings are.

The project investigated how AI is experienced by people who work with it every day. Through ethnographic fieldwork, interviews with caseworkers, and participatory design workshops, the project explored what happens when welfare institutions adopt AI-assisted tools in decision-making. Caseworkers identified the moment of discretion, adapting laws to complex citizen cases, as what makes their work meaningful, and something that must not be compromised by algorithmic solutions. Instead, they welcomed AI support for tasks such as navigating scattered documents and translations. Understanding decision-making contexts, the research argues, is essential to proposing AI solutions that support caseworkers' professionalism.
The project also examined the EU AI Act's implications for innovation and accountability, and explored regulatory sandboxes as tools for responsible AI development, asking how civil society might play a more active role in shaping these frameworks.

Takaisin apurahalistaukseen