論文アブストラクト：Algorithms are increasingly being incorporated into diverse services that orchestrate multiple stakeholders' needs and interests. How can we design these algorithmic services to make decisions that are not only efficient, but also fair and motivating? We take a human-centered approach to identify and address challenges in building human-centered algorithmic services. We are in the process of building an allocation algorithm for 412 Food Rescue, an organization that matches food donations with non-profit organizations. As part of this ongoing project, we conducted interviews with multiple stakeholders in the service-organization staff, donors, volunteers, recipient non-profits and their clients, and everyday citizens-in order to understand how the allocation algorithm, interfaces, and surrounding work practices should be designed. The findings suggest that we need to understand and account for varying fairness notions held by stakeholders; consider people, contexts, and interfaces for algorithms to work fairly in the real world; and preserve meaningfulness and social interaction in automation in order to build fair and motivating algorithmic services.