Experts in the Shadow of Algorithmic Systems

Algorithms support decision-making in various contexts, often diminishing human agency in the process. Without meaningful human input, use of predictive systems can result in costly errors, leaving users unable to evaluate accuracy. Intelligibility is one design criterion that may ensure users remain in the decision-making loop. However, guidance is currently diffuse and focused on the lay user, ignoring the role of expertise. We propose a cognitive psychology-based framework that segments decision-making space by users' expertise, risk-environment and motivation. We illustrate this by focusing on expertise, exploring how we might inform usable intelligibility in interface design, enhancing user agency in the decision-making process.