Elsa labs as method to study AI-based systems at the micro, meso, and macro level

Because existing frameworks in the ethics of AI often miss socio-ethical aspects of AI beyond the micro-level concerns, while structural impacts at meso and macro level raise societal concerns, there is a theoretical and empirical gap in research on AI’s multi-level impacts, and even fewer approaches that can translate insights across these levels into concrete design practices and policy interventions. This chapter introduces the ELSA Lab methodology, aparticipatory, multi-level approach from the Netherlands that engages Quadruple Helix stakeholders(researchers, companies, civilsociety, and policymakers) to broaden the perspectives on responsible AI. Unlike existing frameworks, ELSA Labs address not only micro-level issues but also meso- and macro- level considerations. The chapter outlines the problem and limitations of current frameworks (Sect.2), presents a state-of-the-art overview of ELSA research (Sect.3), introduces the ELSA Lab methodology with its principles and interventions (Sect.4), and concludes with a discussion of its advantages, limitations, and directions for future work.