Redan 1956 skämtade en tysk tecknare om framtidens helautomatiserade förvaltning. De mänskliga byråkraterna skulle hamna på museum och ersättas av en effektiv supermaskin. Sjuttio år senare är skämtet på väg att bli verklighet. AI-byråkratin utlovar större effektivitet, bättre samhällsservice och mer objektiva beslut. Men vad händer när den artificiella intelligensen flyttar in i välfärdsstatens maskinrum? När vi människor möter Försäkringskassans digitala plattformar, Arbetsförmedlingens chattbot Molly, eller handläggarroboten i Trelleborgs kommun? Är AI verkligen en bättre byråkrat? Eller byter vi bara bristfällig mänsklighet mot exakt blindhet när vi lämnar Max Webers järnbur för algoritmernas svarta låda?

This chapter presents the findings of an extensive mapping exercise to systematically identify systems for automated decision-making including aritificial intelligence for welfare provision in Swedisch and Danish municipalities and regions. The starting point for the mapping exercise and the chapter is that knowing the extent and character of such systems is crucial for supporting and maintaining social justice. Theoretically, the contribution is framed through references to discussions of ADM and AI transparency and accountability. While considering the structural and practical limitations of transparency, we end the chapter with suggestions for future research to make welfare automation visible and hence open up for civic participation.

Katarina Winter och Anne Kaun utvecklar i den här artikeln begreppet ”digitala trygghetsteknologier” för att beskriva nya digitala initiativ för att öka tryggheten i våra offentliga rum. Med avstamp i forskning om säkerhetisering och i teknik- och vetenskapsstudier (STS) visar de att även till synes ”mjuka” teknologier, som ofta presenteras som mindre ingripande och mer legitima, riskerar att normalisera och förstärka en logik där trygghet görs till en fråga om övervakning, kartläggning och kontroll. Den begreppsliga diskussionen illustreras empiriskt med material från tre fallstudier av olika typer av fenomen som faller under deras begrepp. Genom att introducera och problematisera begreppet digitala trygghetsteknologier vill Winter och Kaun möjliggöra analyser av teknologiers diskursiva, materiella och politiska implikationer för hur trygghet produceras, för vilka subjekt som målas upp som hot eller skyddsvärda, samt för hur ansvar förskjuts mellan offentliga och privata aktörer.

Artist Verena Friedrich, assistant professor at the Academy of Media Arts Cologne, Germany presented her installation EZ Quality Sorter V2 as part of the ERBSENZÄHLER project, exploring the increasing quantification of life. The event, introduced by Professor Anne Kaun, professor of media and communication studies at Södertörn University, Sweden, delved into themes of classification, objectivity, and the human/technology relationship.

Dr. Alison Powell, Associate Professor in the Department of Media and Communications at London School of Economics, joined the discussion, highlighting how art can bridge gaps between artists, scientists, and the public. And further dialogue was conducted on how the installation challenges societal norms like automation bias, provoking thought on the neutrality of technology.

Through combining art, design, and engineering, Friedrich’s work invites us to reflect on our interactions with technology and what effect it has on our relationship with nature. 

This country report gives an overview of the landscape of automated decision-makingsystems (ADM) within the German public welfare sector. The aim of this working paper is todocument the desk research process that served as a basis for the case study selectionwithin to the Automating Welfare (AUTO-WELF) project. Next to the overview,characteristics of the German ADM landscape derived from the search results arepresented. The desk research was carried out between January and April 2023, with anupdate in January 2024, and includes information on ADM systems until this point in time.The report focuses on ADM projects on a federal, state, and regional level.

This report aims to give a first national overview of automated decision-making systems - ADM -adopted in the public sector in Italy. In the first part of the report it is intended to take a look at whatmethod was adopted for the research and the context in which this research was conducted. Then, inthe second part of the report, an attempt was made to highlight where ADM cases have beenimplemented in the Italian public administration, focusing on welfare sectors (core welfare sector)and municipal sectors (smart city projects). The aim is to give a framing in descriptive terms of theItalian situation in general. Finally, in the last part, four potential case studies, inherent to the field,are described, attempting to detect the potential ADM systems in use or in the making and theirpossible critical points.

This report summarizes the findings from desk research in Sweden and Estonia mapping and documenting automated decision-making tools and support systems in use in the public sector of both countries. The report provides a general overview over governance and implementation principles as well as several shorter descriptions of specific case studies in both countries. The findings indicate that there are several ADM applications in use in both the Swedish and Estonian public sector that range from administrative backend automation to front end automation of communication channels including chatbots. The landscape is however fast changing partly because of the fast-evolving technology but also and especially due to the primarily pilot character of many ADM projects in both countries. The report ends by outlining methodological challenges and provides future directions for further research.