Democratic Participation and Artificial Intelligence in Living Labs: Developing a Feminist Framework to Move Beyond Patriarchy

Democratic systems around the world are under pressure: in addition to political forces on the far right, the deficiencies of liberal models of democracy are subject to scientific debate. A common response to this is the active inclusion of citizens in the political system. Experiments such as living labs are considered a viable strategy to address many kinds of challenges governance arrangements struggle to process – supposedly more democratically, at low costs, and with minimal risks. In this context, AI-driven assistance systems are gaining momentum as it is assumed that these can support processes even more efficiently. Given the often intuitive user interface of many applications, they are already being used extensively without considering the reproduction of social inequalities and stereotypes, and to what extent the development and provision of AI methods might restructure the distribution of political power. Informed by a feminist perspective, the paper proposal at hand then asks: (1) What are the individual, institutional and technological capabilities needed to democratically choose, further developed and apply AI methods in experimental participatory processes and (2) how can these capabilities be developed? Experiments such as real-world or living labs aim at strengthening representation of the people’s will and the democratic system’s overall resilience to multi-crises. Usually linked to normative orientations such as sustainable development, these formats follow a multi-phase cycle, including contextualization, issue selection, concretization, implementation, and feedback. This way, the living lab aims at iterative learning. Experiments can help to identify effective strategies before scaling them up or institutionalizing them. Moreover, such experiments can revitalize stagnant governance processes, open new thematic fields, and strengthen local in-novation ecosystems. AI-driven methods are considered to be able to support these complex decision-making processes by recognizing and understanding heterogeneous user needs and systemically improving performance for dynamic, changing contexts. As mentioned, these opportunities are not without risks: legal uncertainty, bias in training data or decisions, erroneous results due to human decision-makers, a lack of transparency, and the control of technologies by powerful corporations (or individuals) that have little or no democratic legitimacy. The proposed paper’s aim is to identify individual, institutional and technological capabilities needed for integrating AI methods more democratically in participatory processes. Technological capabilities are of particular interest, considering that AI methods are increasingly being integrated as interactive companions in (democratic) processes. Accordingly, the open question of what status AI methods have as actors in democratic processes and how they interact with individuals and institutions must be clarified. A feminist perspective is of interest here, as it is able to critically question powerful social norms that naturalize truths and thus contribute to the sedimentation of problematic conditions. At the same time, such an analysis reveals alternatives in the production and use of digital technologies. The paper will review existing literature on the use of AI methods in different participatory contexts and propose a framework for feminist capabilities for AI-driven democratic participation. Such a framework can support analyses and planning of participatory formats that actually strengthen democratic systems, potentially moving beyond patriarchy.