REPAIR approaches algorithmic systems with three interconnected lenses 

REPAIR uses the concept of 'breakage' to describe situations revealing algorithmic systems' societal interdependencies and infrastructural vulnerabilities. REPAIR assumes that the breakages of algorithmic systems are not solely technological but are also related to how the systems implement or erode trust. The project studies how privacy and trust are ensured when public health interventions are delegated to apps and how organizations respond to real and potential breakages of algorithmic systems. Moreover, it examines infrastructural vulnerabilities, develops ways to improve algorithmic systems and studies how cyber insurance responds to the need to manage datafied uncertainty.

REPAIR addresses conceptual and regulatory questions that need to be considered when repairing the algorithmic system. That requires understanding the technology in the context of social practices and institutions. REPAIR explores how regulatory mechanisms could facilitate the repair work of algorithmic systems. Moreover, it advances the value-sensitive design approaches, promoting concepts and tools for designing and implementing algorithmic systems in a societally sustainable, trust-enhancing manner.

While exploring how trust and security emerge in technology-mediated practices, REPAIR seeks routes for a fair and responsible digital society. That requires identifying the particular situations that combine the strengths of machines with those of humans and learning to make the most of them. REPAIR studies how trust can be ensured when organizations adopt AI-based tools that facilitate or require new forms of human–machine collaboration. By exploring alternative models of human-machine cooperation in social work, healthcare, AI training and knowledge work, REPAIR will create visions for a digital society where trust can be maintained and renewed while designing and deploying algorithmic systems in different societal sectors.