Invisible Humans in AI Governance: Citizenship, Datafication, and Regional AI Strategies in Sweden
- Date
- Thursday 29 January 2026
- Time
- Location
- Kamerlingh Onnes Building
Steenschuur 25
2311 ES Leiden - Room
- A051 - Grotius room
As regulators design new policies and regulatory frameworks for artificial intelligence, how do they conceptualize citizens and account for the ways in which AI affects their rights and lived experiences? Are citizens, as diverse human beings, meaningfully integrated into contemporary AI regulatory frameworks, or are they treated primarily as variables within the logical and mathematical rationales that have long shaped AI research? This question becomes particularly salient when AI systems are deployed in deeply human contexts such as social welfare, which do not operate solely according to efficiency-driven logics but are instead grounded in social values, solidarity, and inclusion. This Article examines these tensions through an analysis of the datafication of the Swedish social welfare system, often regarded as a paradigmatic model of solidarity and citizen-centric social security.
At first glance, a striking mismatch emerges between the marginal presence of citizens in AI policy frameworks and the foundational pillars of the Scandinavian welfare model. This mismatch, however, is neither accidental nor purely technological. It is rooted in longer historical processes, including neoliberal governance, New Public Management reforms, politicized narratives of social welfare fraud, and the expansion of datafication policies. Datafication—the quantification of social interactions, preferences, and characteristics for purposes of analysis, governance, and prediction—now underpins the use of AI systems in the public sector, including contemporary digital welfare states. Yet it remains insufficiently examined whether datafication adequately reflects the distinctive features and normative commitments of welfare states such as Sweden, or whether it instead marks a new phase in their ongoing transformation. The paper argues that AI and the broader datafication of social welfare have consolidated a novel approach to social rights in which solidarity, diversity, and empathy toward citizens have become secondary, while efficiency, rationality, and the self-responsibilisation of individuals for their own destinies have become central.
To examine how this shift manifests in practice, this interdisciplinary and empirical Article analyzes regional public sector AI strategies in Sweden, focusing on how these strategies conceptualize, datafy, and govern citizens within the welfare state. The regional level provides a critical site of inquiry, given the central role Swedish regions play in the delivery of welfare services and their proximity to citizens in everyday administrative interactions. Drawing on a practice-oriented document analysis, the Article examines how citizens are represented in regional AI strategies and how their needs, vulnerabilities, and rights are translated into data-driven governance logics. By foregrounding regional AI governance—an underexplored level in existing scholarship—the Article contributes to debates on AI regulation, digital welfare states, and datafication, while offering broader insights into the ongoing transformation of the Swedish welfare model and the risks posed to citizen-centered welfare governance.
The title and abstract are from a working paper, by Ellinor Blom Lussi, Lund University, Sweden & Sofia Ranchordás, Tilburg University, The Netherlands.
About the speaker
Sofia Ranchordas has been a professor of administrative law at Tilburg Law School since December 2022. Between 2023 and 2028, Ranchordas will be working on her NWO-Vidi project on vulnerability in the automated state. This project addresses the core of administrative law in the digital age: the (unequal) relationship between citizens and administrative bodies and how technology reinforces existing inequalities. With the chair in administrative law at Tilburg University, she aims to generate more attention for the "modern" aspects of this legal field and the pressing problems that vulnerable citizens, such as the victims of the Child Benefits scandal (Toeslagenaffaire), experience daily in their interactions with the government. She teaches administrative law (Dutch and comparative administrative law). Ranchordas is also a part-time professor of public law, innovation, and sustainability at Luiss Guido Carli, where she teaches courses on regulation and innovation policy.